{ "metadata": { "total_size": 22366403936 }, "weight_map": { "talker.codec_head.weight": "model-00005-of-00005.safetensors", "talker.model.embed_tokens.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.0.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.0.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.0.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.0.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.1.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.1.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.1.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.1.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.10.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.10.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.10.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.10.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.11.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.11.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.11.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.11.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.12.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.12.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.12.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.12.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.13.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.13.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.13.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.13.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.14.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.14.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.14.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.14.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.15.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.15.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.15.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.15.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.16.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.16.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.16.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.16.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.17.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.17.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.17.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.17.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.18.input_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.18.mlp.down_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.18.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.18.mlp.up_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.18.post_attention_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.18.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.18.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.18.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.18.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.18.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.18.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.18.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.19.input_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.mlp.down_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.mlp.gate_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.mlp.up_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.post_attention_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.self_attn.k_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.19.self_attn.k_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.self_attn.o_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.self_attn.q_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.19.self_attn.q_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.19.self_attn.v_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.19.self_attn.v_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.2.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.2.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.2.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.2.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.2.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.20.input_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.mlp.down_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.mlp.gate_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.mlp.up_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.post_attention_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.self_attn.k_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.20.self_attn.k_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.self_attn.o_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.self_attn.q_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.20.self_attn.q_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.20.self_attn.v_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.20.self_attn.v_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.input_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.mlp.down_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.mlp.gate_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.mlp.up_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.post_attention_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.self_attn.k_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.21.self_attn.k_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.self_attn.o_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.self_attn.q_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.21.self_attn.q_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.21.self_attn.v_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.21.self_attn.v_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.input_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.mlp.down_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.mlp.gate_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.mlp.up_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.post_attention_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.self_attn.k_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.22.self_attn.k_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.self_attn.o_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.self_attn.q_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.22.self_attn.q_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.22.self_attn.v_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.22.self_attn.v_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.input_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.mlp.down_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.mlp.gate_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.mlp.up_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.post_attention_layernorm.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.self_attn.k_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.23.self_attn.k_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.self_attn.o_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.self_attn.q_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.23.self_attn.q_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.23.self_attn.v_proj.bias": "model-00005-of-00005.safetensors", "talker.model.layers.23.self_attn.v_proj.weight": "model-00005-of-00005.safetensors", "talker.model.layers.3.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.3.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.3.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.3.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.3.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.4.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.4.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.4.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.4.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.5.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.5.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.5.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.5.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.6.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.6.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.6.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.6.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.7.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.7.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.7.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.7.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.8.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.8.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.8.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.8.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.input_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.9.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.9.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "talker.model.layers.9.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "talker.model.layers.9.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "talker.model.norm.weight": "model-00005-of-00005.safetensors", "talker.thinker_to_talker_proj.bias": "model-00004-of-00005.safetensors", "talker.thinker_to_talker_proj.weight": "model-00004-of-00005.safetensors", "thinker.audio_tower.audio_bos_eos_token.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.conv1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.conv1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.conv2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.conv2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.0.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.1.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.10.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.11.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.12.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.13.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.14.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.15.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.16.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.17.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.18.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.19.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.2.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.20.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.21.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.22.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.23.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.24.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.25.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.26.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.27.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.28.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.29.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.3.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.30.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.31.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.4.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.5.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.6.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.7.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.8.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.fc1.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.fc1.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.fc2.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.fc2.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.final_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.final_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn.out_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn.out_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn_layer_norm.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.layers.9.self_attn_layer_norm.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.ln_post.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.ln_post.weight": "model-00001-of-00005.safetensors", "thinker.audio_tower.proj.bias": "model-00001-of-00005.safetensors", "thinker.audio_tower.proj.weight": "model-00001-of-00005.safetensors", "thinker.lm_head.weight": "model-00004-of-00005.safetensors", "thinker.model.embed_tokens.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.input_layernorm.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.self_attn.k_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.0.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.input_layernorm.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.self_attn.k_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.1.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.10.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.10.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.10.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.10.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.10.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.11.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.11.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.11.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.11.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.12.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.12.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.12.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.12.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.13.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.13.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.13.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.13.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.13.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.13.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.13.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.13.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.13.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.13.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.13.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.13.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.14.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.14.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.14.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.14.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.14.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.15.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.15.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.15.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.15.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.16.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.16.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.16.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.16.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.17.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.17.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.17.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.17.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.18.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.18.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.18.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.18.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.19.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.19.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.19.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.19.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.2.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.2.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.2.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.2.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.2.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.2.self_attn.k_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.2.self_attn.q_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.2.self_attn.v_proj.bias": "model-00001-of-00005.safetensors", "thinker.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00005.safetensors", "thinker.model.layers.20.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.20.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.20.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.20.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.20.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.21.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.21.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.21.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.21.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.22.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.22.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.22.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.22.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.input_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.mlp.down_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.mlp.gate_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.mlp.up_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.post_attention_layernorm.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.23.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.23.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.23.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.23.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.24.input_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.24.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.24.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.24.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.24.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.24.self_attn.k_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.24.self_attn.k_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.24.self_attn.o_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.24.self_attn.q_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.24.self_attn.q_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.24.self_attn.v_proj.bias": "model-00003-of-00005.safetensors", "thinker.model.layers.24.self_attn.v_proj.weight": "model-00003-of-00005.safetensors", "thinker.model.layers.25.input_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.25.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.25.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.25.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.25.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.input_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.26.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.26.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.26.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.26.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.input_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.mlp.down_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.mlp.gate_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.mlp.up_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.post_attention_layernorm.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.self_attn.k_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.27.self_attn.k_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.self_attn.o_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.self_attn.q_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.27.self_attn.q_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.27.self_attn.v_proj.bias": "model-00004-of-00005.safetensors", "thinker.model.layers.27.self_attn.v_proj.weight": "model-00004-of-00005.safetensors", "thinker.model.layers.3.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.3.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.3.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.3.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.3.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.4.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.4.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.4.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.4.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.5.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.5.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.5.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.5.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.6.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.6.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.6.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.6.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.7.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.7.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.7.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.7.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.8.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.8.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.8.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.8.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.input_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.mlp.down_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.mlp.gate_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.mlp.up_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.post_attention_layernorm.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.self_attn.k_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.9.self_attn.k_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.self_attn.o_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.self_attn.q_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.9.self_attn.q_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.layers.9.self_attn.v_proj.bias": "model-00002-of-00005.safetensors", "thinker.model.layers.9.self_attn.v_proj.weight": "model-00002-of-00005.safetensors", "thinker.model.norm.weight": "model-00004-of-00005.safetensors", "thinker.visual.blocks.0.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.mlp.down_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.mlp.down_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.mlp.gate_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.mlp.gate_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.mlp.up_proj.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.mlp.up_proj.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.norm1.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.norm2.weight": "model-00001-of-00005.safetensors", "thinker.visual.merger.ln_q.weight": "model-00001-of-00005.safetensors", "thinker.visual.merger.mlp.0.bias": "model-00001-of-00005.safetensors", "thinker.visual.merger.mlp.0.weight": "model-00001-of-00005.safetensors", "thinker.visual.merger.mlp.2.bias": "model-00001-of-00005.safetensors", "thinker.visual.merger.mlp.2.weight": "model-00001-of-00005.safetensors", "thinker.visual.patch_embed.proj.weight": "model-00001-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.activation_post.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.activation_post.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.conv_post.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.conv_pre.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.conv_pre.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.0.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.1.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.10.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.11.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.12.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.13.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.14.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.15.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.16.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.17.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.2.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.3.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.4.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.5.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.6.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.7.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.8.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.0.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.0.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.1.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.1.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.2.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.2.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.3.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.3.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.4.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.4.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.5.act.alpha": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.activations.5.act.beta": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs1.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs1.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs1.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs1.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs2.1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs2.1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs2.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.resblocks.9.convs2.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.0.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.0.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.1.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.1.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.2.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.2.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.3.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.3.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.4.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.4.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.5.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_bigvgan_model.ups.5.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.proj.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.proj.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.norm_out.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.norm_out.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.proj_out.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.proj_out.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.rotary_embed.inv_freq": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.text_embed.codec_embed.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.time_embed.time_mlp.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.time_embed.time_mlp.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.time_embed.time_mlp.2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.time_embed.time_mlp.2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_k.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_k.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_out.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_out.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_q.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_q.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_v.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn.to_v.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn_norm.linear.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.attn_norm.linear.weight": "model-00005-of-00005.safetensors", "thinker.visual.blocks.0.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.0.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.1.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.10.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.11.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.12.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.13.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.14.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.15.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.16.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.17.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.18.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.19.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.2.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.20.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.21.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.22.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.23.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.24.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.25.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.26.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.27.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.28.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.29.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.3.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.30.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.31.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.4.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.5.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.6.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.7.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.8.attn.v.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.q.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.k.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.v.bias": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.q.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.k.weight": "model-00001-of-00005.safetensors", "thinker.visual.blocks.9.attn.v.weight": "model-00001-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.asp.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.asp.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.asp.tdnn.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.asp.tdnn.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.0.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.0.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.res2net_block.blocks.0.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.res2net_block.blocks.0.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.se_block.conv1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.se_block.conv1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.se_block.conv2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.se_block.conv2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.tdnn1.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.tdnn1.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.tdnn2.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.1.tdnn2.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.res2net_block.blocks.0.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.res2net_block.blocks.0.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.se_block.conv1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.se_block.conv1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.se_block.conv2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.se_block.conv2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.tdnn1.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.tdnn1.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.tdnn2.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.2.tdnn2.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.res2net_block.blocks.0.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.res2net_block.blocks.0.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.se_block.conv1.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.se_block.conv1.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.se_block.conv2.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.se_block.conv2.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.tdnn1.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.tdnn1.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.tdnn2.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.blocks.3.tdnn2.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.fc.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.fc.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.mfa.conv.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.input_embed.spk_encoder.mfa.conv.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.0.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.1.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.10.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.11.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.12.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.13.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.14.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.15.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.16.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.17.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.18.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.19.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.2.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.20.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.21.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.3.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.4.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.5.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.6.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.7.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.8.ff.ff.3.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.ff.ff.0.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.ff.ff.0.weight": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.ff.ff.3.bias": "model-00005-of-00005.safetensors", "token2wav.code2wav_dit_model.transformer_blocks.9.ff.ff.3.weight": "model-00005-of-00005.safetensors" } }