|
{ |
|
"architectures": [ |
|
"Doge2ForCausalLM" |
|
], |
|
"attention_bias": false, |
|
"attention_dropout": 0.0, |
|
"auto_map": { |
|
"AutoConfig": "configuration_doge2.Doge2Config", |
|
"AutoModelForCausalLM": "modeling_doge2.Doge2ForCausalLM" |
|
}, |
|
"bos_token_id": 151643, |
|
"dynamic_mask_ratio": 0.0, |
|
"eos_token_id": 151645, |
|
"hidden_act": "silu", |
|
"hidden_dropout": 0.0, |
|
"hidden_size": 2048, |
|
"initializer_range": 0.02, |
|
"intermediate_size": 6144, |
|
"is_moe": false, |
|
"keep_window_size": 8192, |
|
"max_position_embeddings": 16384, |
|
"model_type": "doge2", |
|
"norm_topk_prob": false, |
|
"num_attention_heads": 16, |
|
"num_experts": 4096, |
|
"num_experts_per_tok": 64, |
|
"num_hidden_layers": 28, |
|
"num_key_value_heads": 8, |
|
"output_router_logits": false, |
|
"pad_token_id": 151643, |
|
"rms_norm_eps": 1e-06, |
|
"rope_scaling": null, |
|
"rope_theta": 100000.0, |
|
"router_aux_loss_coef": 0.001, |
|
"tie_word_embeddings": true, |
|
"torch_dtype": "bfloat16", |
|
"transformers_version": "4.52.3", |
|
"unk_token_id": 151643, |
|
"use_cache": true, |
|
"vocab_size": 151936 |
|
} |
|
|