Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

added_tokens.json

@@ -0,0 +1,28 @@
+{
+  "</think>": 151668,
+  "</tool_call>": 151658,
+  "</tool_response>": 151666,
+  "<think>": 151667,
+  "<tool_call>": 151657,
+  "<tool_response>": 151665,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

config.json

@@ -0,0 +1,28 @@
+{
+  "_name_or_path": "./model",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 27648,
+  "max_position_embeddings": 131072,
+  "max_window_layers": 64,
+  "model_type": "qwen2",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 64,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": 151645,
+  "transformers_version": "4.49.0"
+}

latest

@@ -0,0 +1 @@
+global_step379

model-00001-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea4820fb1379c53257f82101456fb80ffa0ad686e7f4d2101be1d82a55e9abe6
+size 4891730992

model-00002-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f10b6678bdd5b0325d0acfd804e01b0307fe4807b60ce1e422021184f8eefa5
+size 4876059352

model-00003-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:803ac3534b7ac607b629e89766f6caf2bd7ce8bda56351934dc9fff7f720500a
+size 4876059384

model-00004-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c5692f2d456b5aab90d86bb4a85c9357cb38d2b40fc8871312e7a25701a2abb
+size 4876059416

model-00005-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e452e1c589940330cd4d4fcc0b3c6ec66541a1e5aecb6b576fa054f4c963024
+size 4876059416

model-00006-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6b2ca471378a8c5dfd740bd5a68244363e171cea845260188c7b9a954d3f081
+size 4876059416

model-00007-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ca93426e11c7bf0499a1d6801f5444417d05c69d7260dbf9209bb8c5e3f2e99
+size 4876059416

model-00008-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8936084885a8742dd8ee719115fc474c2fca7d8d50948a1206ed94d5af3504ef
+size 4876059416

model-00009-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d52772d0dbd0c058bbef4c5b7921f2b6886c9095ead0eee93f3f7a373f536fa
+size 4876059416

model-00010-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:865d9be31577123e6fb70065a7c2e7a6eeedbf780a0fbabe0b09a4494d2060d1
+size 4876059416

model-00011-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c7d9422157230cb63f89ede398edeb93f9aa36f99a8ad3675e7790d31b601
+size 4876059416

model-00012-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6db19a625e1219a276a503f679ac443dad505677fb895f4d119017d3551f594d
+size 4876059416

model-00013-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:712555ee9aecb2aff1895fdddbbe023805691cb126d072b03f81059c35dc3739
+size 4876059416

model-00014-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61cde7c333ac18d631c24377a0cbeb84bd8d4e054e8204ff3241922d25d3132a
+size 2123397800

model.safetensors.index.json

@@ -0,0 +1,778 @@
+{
+  "metadata": {
+    "total_size": 65527752704
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00014-of-00014.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00004-of-00014.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00005-of-00014.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.26.self_attn.v_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.28.self_attn.k_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.28.self_attn.q_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.28.self_attn.v_proj.bias": "model-00006-of-00014.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.self_attn.k_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.self_attn.q_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.29.self_attn.v_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00014.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00014.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.self_attn.k_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.self_attn.q_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.30.self_attn.v_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.self_attn.k_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.self_attn.q_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.31.self_attn.v_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.input_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.post_attention_layernorm.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.self_attn.k_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.32.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.self_attn.q_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.32.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.32.self_attn.v_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.32.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.33.input_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.33.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.33.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.33.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.33.post_attention_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.33.self_attn.k_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.33.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.33.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.33.self_attn.q_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.33.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.33.self_attn.v_proj.bias": "model-00007-of-00014.safetensors",
+    "model.layers.33.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
+    "model.layers.34.input_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.post_attention_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.self_attn.k_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.34.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.self_attn.q_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.34.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.34.self_attn.v_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.34.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.input_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.post_attention_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.self_attn.k_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.35.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.self_attn.q_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.35.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.35.self_attn.v_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.35.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.input_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.post_attention_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.self_attn.k_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.36.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.self_attn.q_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.36.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.36.self_attn.v_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.36.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.input_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.post_attention_layernorm.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.self_attn.k_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.37.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.self_attn.q_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.37.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.37.self_attn.v_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.37.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.38.input_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.38.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.38.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.38.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.38.post_attention_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.38.self_attn.k_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.38.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.38.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.38.self_attn.q_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.38.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.38.self_attn.v_proj.bias": "model-00008-of-00014.safetensors",
+    "model.layers.38.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
+    "model.layers.39.input_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.post_attention_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.self_attn.k_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.39.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.self_attn.q_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.39.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.39.self_attn.v_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.39.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.40.input_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.post_attention_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.self_attn.k_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.40.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.self_attn.q_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.40.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.40.self_attn.v_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.40.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.input_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.post_attention_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.self_attn.k_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.41.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.self_attn.q_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.41.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.41.self_attn.v_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.41.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.input_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.post_attention_layernorm.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.self_attn.k_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.42.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.self_attn.q_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.42.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.42.self_attn.v_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.42.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.43.input_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.43.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.43.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.43.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.43.post_attention_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.43.self_attn.k_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.43.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.43.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.43.self_attn.q_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.43.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.43.self_attn.v_proj.bias": "model-00009-of-00014.safetensors",
+    "model.layers.43.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
+    "model.layers.44.input_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.post_attention_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.self_attn.k_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.44.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.self_attn.q_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.44.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.44.self_attn.v_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.44.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.input_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.post_attention_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.self_attn.k_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.45.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.self_attn.q_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.45.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.45.self_attn.v_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.45.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.input_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.post_attention_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.self_attn.k_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.46.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.self_attn.q_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.46.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.46.self_attn.v_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.46.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.input_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.post_attention_layernorm.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.self_attn.k_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.47.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.self_attn.q_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.47.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.47.self_attn.v_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.47.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.48.input_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.48.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.48.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.48.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.48.post_attention_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.48.self_attn.k_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.48.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.48.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.48.self_attn.q_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.48.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.48.self_attn.v_proj.bias": "model-00010-of-00014.safetensors",
+    "model.layers.48.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
+    "model.layers.49.input_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.post_attention_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.self_attn.k_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.49.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.self_attn.q_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.49.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.49.self_attn.v_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.49.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.50.input_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.post_attention_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.self_attn.k_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.50.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.self_attn.q_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.50.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.50.self_attn.v_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.50.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.input_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.post_attention_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.self_attn.k_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.51.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.self_attn.q_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.51.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.51.self_attn.v_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.51.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.input_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.post_attention_layernorm.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.self_attn.k_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.52.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.self_attn.q_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.52.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.52.self_attn.v_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.52.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.53.input_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.53.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.53.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.53.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.53.post_attention_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.53.self_attn.k_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.53.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.53.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.53.self_attn.q_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.53.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.53.self_attn.v_proj.bias": "model-00011-of-00014.safetensors",
+    "model.layers.53.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
+    "model.layers.54.input_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.post_attention_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.self_attn.k_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.54.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.self_attn.q_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.54.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.54.self_attn.v_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.54.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.input_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.post_attention_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.self_attn.k_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.55.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.self_attn.q_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.55.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.55.self_attn.v_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.55.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.input_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.post_attention_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.self_attn.k_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.56.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.self_attn.q_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.56.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.56.self_attn.v_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.56.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.input_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.post_attention_layernorm.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.self_attn.k_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.57.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.self_attn.q_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.57.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.57.self_attn.v_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.57.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.58.input_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.58.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.58.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.58.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.58.post_attention_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.58.self_attn.k_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.58.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.58.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.58.self_attn.q_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.58.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.58.self_attn.v_proj.bias": "model-00012-of-00014.safetensors",
+    "model.layers.58.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
+    "model.layers.59.input_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.post_attention_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.self_attn.k_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.59.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.self_attn.q_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.59.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.59.self_attn.v_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.59.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.60.input_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.post_attention_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.self_attn.k_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.60.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.self_attn.q_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.60.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.60.self_attn.v_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.60.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.input_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.post_attention_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.self_attn.k_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.61.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.self_attn.q_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.61.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.61.self_attn.v_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.61.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.input_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.post_attention_layernorm.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.self_attn.k_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.62.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.self_attn.q_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.62.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.62.self_attn.v_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.62.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.63.input_layernorm.weight": "model-00014-of-00014.safetensors",
+    "model.layers.63.mlp.down_proj.weight": "model-00014-of-00014.safetensors",
+    "model.layers.63.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.63.mlp.up_proj.weight": "model-00014-of-00014.safetensors",
+    "model.layers.63.post_attention_layernorm.weight": "model-00014-of-00014.safetensors",
+    "model.layers.63.self_attn.k_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.63.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.63.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.63.self_attn.q_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.63.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.63.self_attn.v_proj.bias": "model-00013-of-00014.safetensors",
+    "model.layers.63.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00002-of-00014.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00003-of-00014.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
+    "model.norm.weight": "model-00014-of-00014.safetensors"
+  }
+}

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85
+size 15984

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f338ce80d7c441076bfc8c53b84067a0181f5a14e80c13d5acb8150b659f4d73
+size 15984

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fbc9fa428939be10b46779f0eb5cd833e0da426b1cbdee77b3a55b6952235b
+size 15984

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac55dba0b79d5fa4699d239da2f966d52040d576d31234ac8d4632e6956481bc
+size 15984

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af2d0c015100768ffa23faf3b6c2d54ea89eb045603e30e55cd211e06ff34972
+size 15984

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c60a1b40608e34bc801c8231f97b81c53b5290dfaed1b9cd0ccbeca29574a991
+size 15984

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ad6a142a403eb9aafc4a3a9a856bca648fe31fd22d796867baca31fb13656aa
+size 15984

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38bc23a138cc800b22881742c0f3f9a71731a9a7111c6058a0077e6274d21773
+size 15984

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c15f1bc7297a28645464f9c0f3ff75c45e9552cff93fa6b52280fe689c6153ee
+size 1064

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

tokenizer_config.json

@@ -0,0 +1,240 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- '' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n  {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" and not message.tool_calls %}\n        {%- set content = message.content.split('</think>')[-1].lstrip('\\n') %}\n        {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {%- set content = message.content.split('</think>')[-1].lstrip('\\n') %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n<think>\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,2693 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9921259842519685,
+  "eval_steps": 500,
+  "global_step": 380,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.005249343832020997,
+      "grad_norm": 4.262586140207107,
+      "learning_rate": 1.2500000000000002e-07,
+      "loss": 1.2143,
+      "step": 1
+    },
+    {
+      "epoch": 0.010498687664041995,
+      "grad_norm": 4.1559742669756154,
+      "learning_rate": 2.5000000000000004e-07,
+      "loss": 1.2307,
+      "step": 2
+    },
+    {
+      "epoch": 0.015748031496062992,
+      "grad_norm": 4.2196647284049895,
+      "learning_rate": 3.75e-07,
+      "loss": 1.2286,
+      "step": 3
+    },
+    {
+      "epoch": 0.02099737532808399,
+      "grad_norm": 4.13634077943981,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 1.2002,
+      "step": 4
+    },
+    {
+      "epoch": 0.026246719160104987,
+      "grad_norm": 4.015668455829927,
+      "learning_rate": 6.25e-07,
+      "loss": 1.1672,
+      "step": 5
+    },
+    {
+      "epoch": 0.031496062992125984,
+      "grad_norm": 3.832855314884781,
+      "learning_rate": 7.5e-07,
+      "loss": 1.1993,
+      "step": 6
+    },
+    {
+      "epoch": 0.03674540682414698,
+      "grad_norm": 3.8323407788221733,
+      "learning_rate": 8.75e-07,
+      "loss": 1.1554,
+      "step": 7
+    },
+    {
+      "epoch": 0.04199475065616798,
+      "grad_norm": 3.7465244180174917,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 1.1672,
+      "step": 8
+    },
+    {
+      "epoch": 0.047244094488188976,
+      "grad_norm": 3.7827251172961986,
+      "learning_rate": 1.125e-06,
+      "loss": 1.1755,
+      "step": 9
+    },
+    {
+      "epoch": 0.05249343832020997,
+      "grad_norm": 3.470602675526565,
+      "learning_rate": 1.25e-06,
+      "loss": 1.1419,
+      "step": 10
+    },
+    {
+      "epoch": 0.05774278215223097,
+      "grad_norm": 3.556221853917274,
+      "learning_rate": 1.3750000000000002e-06,
+      "loss": 1.194,
+      "step": 11
+    },
+    {
+      "epoch": 0.06299212598425197,
+      "grad_norm": 3.324934060085957,
+      "learning_rate": 1.5e-06,
+      "loss": 1.1336,
+      "step": 12
+    },
+    {
+      "epoch": 0.06824146981627296,
+      "grad_norm": 2.965981688480075,
+      "learning_rate": 1.6250000000000001e-06,
+      "loss": 1.1349,
+      "step": 13
+    },
+    {
+      "epoch": 0.07349081364829396,
+      "grad_norm": 2.8658973663115046,
+      "learning_rate": 1.75e-06,
+      "loss": 1.1776,
+      "step": 14
+    },
+    {
+      "epoch": 0.07874015748031496,
+      "grad_norm": 2.720689909744523,
+      "learning_rate": 1.8750000000000003e-06,
+      "loss": 1.1549,
+      "step": 15
+    },
+    {
+      "epoch": 0.08398950131233596,
+      "grad_norm": 2.439062154183451,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.141,
+      "step": 16
+    },
+    {
+      "epoch": 0.08923884514435695,
+      "grad_norm": 2.1353279033918002,
+      "learning_rate": 2.125e-06,
+      "loss": 1.139,
+      "step": 17
+    },
+    {
+      "epoch": 0.09448818897637795,
+      "grad_norm": 1.2092875477650313,
+      "learning_rate": 2.25e-06,
+      "loss": 1.0516,
+      "step": 18
+    },
+    {
+      "epoch": 0.09973753280839895,
+      "grad_norm": 1.0763847439987342,
+      "learning_rate": 2.375e-06,
+      "loss": 1.0802,
+      "step": 19
+    },
+    {
+      "epoch": 0.10498687664041995,
+      "grad_norm": 1.0340250902529846,
+      "learning_rate": 2.5e-06,
+      "loss": 1.0607,
+      "step": 20
+    },
+    {
+      "epoch": 0.11023622047244094,
+      "grad_norm": 0.8630354040462489,
+      "learning_rate": 2.6250000000000003e-06,
+      "loss": 1.0496,
+      "step": 21
+    },
+    {
+      "epoch": 0.11548556430446194,
+      "grad_norm": 0.8072735722523627,
+      "learning_rate": 2.7500000000000004e-06,
+      "loss": 1.08,
+      "step": 22
+    },
+    {
+      "epoch": 0.12073490813648294,
+      "grad_norm": 0.7305262800316248,
+      "learning_rate": 2.875e-06,
+      "loss": 1.0539,
+      "step": 23
+    },
+    {
+      "epoch": 0.12598425196850394,
+      "grad_norm": 0.7406394208995156,
+      "learning_rate": 3e-06,
+      "loss": 1.0529,
+      "step": 24
+    },
+    {
+      "epoch": 0.13123359580052493,
+      "grad_norm": 0.7903255346265977,
+      "learning_rate": 3.125e-06,
+      "loss": 0.9914,
+      "step": 25
+    },
+    {
+      "epoch": 0.13648293963254593,
+      "grad_norm": 0.8017049075586423,
+      "learning_rate": 3.2500000000000002e-06,
+      "loss": 1.0313,
+      "step": 26
+    },
+    {
+      "epoch": 0.14173228346456693,
+      "grad_norm": 0.7764417012146556,
+      "learning_rate": 3.3750000000000003e-06,
+      "loss": 1.0422,
+      "step": 27
+    },
+    {
+      "epoch": 0.14698162729658792,
+      "grad_norm": 0.7112571206699242,
+      "learning_rate": 3.5e-06,
+      "loss": 1.0187,
+      "step": 28
+    },
+    {
+      "epoch": 0.15223097112860892,
+      "grad_norm": 0.6069574236656299,
+      "learning_rate": 3.625e-06,
+      "loss": 0.9958,
+      "step": 29
+    },
+    {
+      "epoch": 0.15748031496062992,
+      "grad_norm": 0.5747553586770598,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.9858,
+      "step": 30
+    },
+    {
+      "epoch": 0.16272965879265092,
+      "grad_norm": 0.590510044443011,
+      "learning_rate": 3.875e-06,
+      "loss": 0.9841,
+      "step": 31
+    },
+    {
+      "epoch": 0.1679790026246719,
+      "grad_norm": 0.5240510507345066,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.0171,
+      "step": 32
+    },
+    {
+      "epoch": 0.1732283464566929,
+      "grad_norm": 0.4913378134702597,
+      "learning_rate": 4.125e-06,
+      "loss": 1.0218,
+      "step": 33
+    },
+    {
+      "epoch": 0.1784776902887139,
+      "grad_norm": 0.47909247638813307,
+      "learning_rate": 4.25e-06,
+      "loss": 1.02,
+      "step": 34
+    },
+    {
+      "epoch": 0.1837270341207349,
+      "grad_norm": 0.4949722185401137,
+      "learning_rate": 4.3750000000000005e-06,
+      "loss": 1.0206,
+      "step": 35
+    },
+    {
+      "epoch": 0.1889763779527559,
+      "grad_norm": 0.48456911030733396,
+      "learning_rate": 4.5e-06,
+      "loss": 1.0321,
+      "step": 36
+    },
+    {
+      "epoch": 0.1942257217847769,
+      "grad_norm": 0.49965866704889284,
+      "learning_rate": 4.625000000000001e-06,
+      "loss": 1.0184,
+      "step": 37
+    },
+    {
+      "epoch": 0.1994750656167979,
+      "grad_norm": 0.4898360279427832,
+      "learning_rate": 4.75e-06,
+      "loss": 1.0165,
+      "step": 38
+    },
+    {
+      "epoch": 0.2047244094488189,
+      "grad_norm": 0.5039246597121783,
+      "learning_rate": 4.875e-06,
+      "loss": 0.9811,
+      "step": 39
+    },
+    {
+      "epoch": 0.2099737532808399,
+      "grad_norm": 0.462371012430454,
+      "learning_rate": 5e-06,
+      "loss": 0.9999,
+      "step": 40
+    },
+    {
+      "epoch": 0.2152230971128609,
+      "grad_norm": 0.4564370676360458,
+      "learning_rate": 4.99989327925842e-06,
+      "loss": 1.0056,
+      "step": 41
+    },
+    {
+      "epoch": 0.2204724409448819,
+      "grad_norm": 0.4838548423542603,
+      "learning_rate": 4.999573126145132e-06,
+      "loss": 1.0099,
+      "step": 42
+    },
+    {
+      "epoch": 0.22572178477690288,
+      "grad_norm": 0.47657399375831033,
+      "learning_rate": 4.999039567993719e-06,
+      "loss": 1.0059,
+      "step": 43
+    },
+    {
+      "epoch": 0.23097112860892388,
+      "grad_norm": 0.4609228968128241,
+      "learning_rate": 4.998292650357558e-06,
+      "loss": 0.9613,
+      "step": 44
+    },
+    {
+      "epoch": 0.23622047244094488,
+      "grad_norm": 0.5379441335347738,
+      "learning_rate": 4.997332437005932e-06,
+      "loss": 0.9912,
+      "step": 45
+    },
+    {
+      "epoch": 0.24146981627296588,
+      "grad_norm": 0.4488276134846175,
+      "learning_rate": 4.996159009918586e-06,
+      "loss": 0.9623,
+      "step": 46
+    },
+    {
+      "epoch": 0.24671916010498687,
+      "grad_norm": 0.7899879742350473,
+      "learning_rate": 4.994772469278726e-06,
+      "loss": 0.9373,
+      "step": 47
+    },
+    {
+      "epoch": 0.25196850393700787,
+      "grad_norm": 0.45351107525432893,
+      "learning_rate": 4.99317293346447e-06,
+      "loss": 0.9312,
+      "step": 48
+    },
+    {
+      "epoch": 0.2572178477690289,
+      "grad_norm": 0.4145057356223518,
+      "learning_rate": 4.991360539038737e-06,
+      "loss": 0.9133,
+      "step": 49
+    },
+    {
+      "epoch": 0.26246719160104987,
+      "grad_norm": 0.4371226648356658,
+      "learning_rate": 4.989335440737587e-06,
+      "loss": 0.9763,
+      "step": 50
+    },
+    {
+      "epoch": 0.2677165354330709,
+      "grad_norm": 0.49987252416920314,
+      "learning_rate": 4.987097811457015e-06,
+      "loss": 0.9753,
+      "step": 51
+    },
+    {
+      "epoch": 0.27296587926509186,
+      "grad_norm": 0.4737378597917066,
+      "learning_rate": 4.984647842238185e-06,
+      "loss": 0.9509,
+      "step": 52
+    },
+    {
+      "epoch": 0.2782152230971129,
+      "grad_norm": 0.4803218242006868,
+      "learning_rate": 4.981985742251123e-06,
+      "loss": 1.0008,
+      "step": 53
+    },
+    {
+      "epoch": 0.28346456692913385,
+      "grad_norm": 0.4238793960297473,
+      "learning_rate": 4.9791117387768575e-06,
+      "loss": 1.0024,
+      "step": 54
+    },
+    {
+      "epoch": 0.2887139107611549,
+      "grad_norm": 0.4135698693902407,
+      "learning_rate": 4.976026077188013e-06,
+      "loss": 0.9208,
+      "step": 55
+    },
+    {
+      "epoch": 0.29396325459317585,
+      "grad_norm": 0.49893312330659967,
+      "learning_rate": 4.972729020927866e-06,
+      "loss": 0.9771,
+      "step": 56
+    },
+    {
+      "epoch": 0.2992125984251969,
+      "grad_norm": 0.43878390791709027,
+      "learning_rate": 4.9692208514878445e-06,
+      "loss": 0.937,
+      "step": 57
+    },
+    {
+      "epoch": 0.30446194225721784,
+      "grad_norm": 0.42250251333050837,
+      "learning_rate": 4.965501868383507e-06,
+      "loss": 0.9287,
+      "step": 58
+    },
+    {
+      "epoch": 0.30971128608923887,
+      "grad_norm": 0.42426620742455357,
+      "learning_rate": 4.961572389128959e-06,
+      "loss": 0.9374,
+      "step": 59
+    },
+    {
+      "epoch": 0.31496062992125984,
+      "grad_norm": 0.4622588583575654,
+      "learning_rate": 4.957432749209755e-06,
+      "loss": 0.99,
+      "step": 60
+    },
+    {
+      "epoch": 0.32020997375328086,
+      "grad_norm": 0.4324798534582787,
+      "learning_rate": 4.953083302054247e-06,
+      "loss": 1.0035,
+      "step": 61
+    },
+    {
+      "epoch": 0.32545931758530183,
+      "grad_norm": 0.4303590460079348,
+      "learning_rate": 4.948524419003415e-06,
+      "loss": 0.9585,
+      "step": 62
+    },
+    {
+      "epoch": 0.33070866141732286,
+      "grad_norm": 0.42861048906851473,
+      "learning_rate": 4.943756489279164e-06,
+      "loss": 0.9772,
+      "step": 63
+    },
+    {
+      "epoch": 0.3359580052493438,
+      "grad_norm": 0.4115697149677722,
+      "learning_rate": 4.938779919951092e-06,
+      "loss": 0.9426,
+      "step": 64
+    },
+    {
+      "epoch": 0.34120734908136485,
+      "grad_norm": 0.417300147661056,
+      "learning_rate": 4.933595135901733e-06,
+      "loss": 0.9447,
+      "step": 65
+    },
+    {
+      "epoch": 0.3464566929133858,
+      "grad_norm": 0.4040601468423496,
+      "learning_rate": 4.928202579790285e-06,
+      "loss": 0.966,
+      "step": 66
+    },
+    {
+      "epoch": 0.35170603674540685,
+      "grad_norm": 0.3677161548087925,
+      "learning_rate": 4.9226027120148195e-06,
+      "loss": 0.941,
+      "step": 67
+    },
+    {
+      "epoch": 0.3569553805774278,
+      "grad_norm": 0.3832286188469758,
+      "learning_rate": 4.916796010672969e-06,
+      "loss": 0.9822,
+      "step": 68
+    },
+    {
+      "epoch": 0.36220472440944884,
+      "grad_norm": 0.4345291089971557,
+      "learning_rate": 4.910782971521112e-06,
+      "loss": 0.9687,
+      "step": 69
+    },
+    {
+      "epoch": 0.3674540682414698,
+      "grad_norm": 0.4133843702108161,
+      "learning_rate": 4.904564107932048e-06,
+      "loss": 0.9283,
+      "step": 70
+    },
+    {
+      "epoch": 0.37270341207349084,
+      "grad_norm": 0.38465081625519515,
+      "learning_rate": 4.898139950851163e-06,
+      "loss": 0.9479,
+      "step": 71
+    },
+    {
+      "epoch": 0.3779527559055118,
+      "grad_norm": 0.424340283488392,
+      "learning_rate": 4.891511048751102e-06,
+      "loss": 0.9593,
+      "step": 72
+    },
+    {
+      "epoch": 0.38320209973753283,
+      "grad_norm": 0.40638719244646565,
+      "learning_rate": 4.884677967584945e-06,
+      "loss": 0.9264,
+      "step": 73
+    },
+    {
+      "epoch": 0.3884514435695538,
+      "grad_norm": 0.4168684939745582,
+      "learning_rate": 4.8776412907378845e-06,
+      "loss": 0.9399,
+      "step": 74
+    },
+    {
+      "epoch": 0.3937007874015748,
+      "grad_norm": 0.3928820462579945,
+      "learning_rate": 4.870401618977415e-06,
+      "loss": 0.9736,
+      "step": 75
+    },
+    {
+      "epoch": 0.3989501312335958,
+      "grad_norm": 0.48005159652101537,
+      "learning_rate": 4.86295957040205e-06,
+      "loss": 0.9935,
+      "step": 76
+    },
+    {
+      "epoch": 0.4041994750656168,
+      "grad_norm": 0.4229896582271501,
+      "learning_rate": 4.855315780388541e-06,
+      "loss": 0.9358,
+      "step": 77
+    },
+    {
+      "epoch": 0.4094488188976378,
+      "grad_norm": 0.4258651800976577,
+      "learning_rate": 4.847470901537642e-06,
+      "loss": 0.9583,
+      "step": 78
+    },
+    {
+      "epoch": 0.4146981627296588,
+      "grad_norm": 0.4069975102427097,
+      "learning_rate": 4.839425603618382e-06,
+      "loss": 0.9237,
+      "step": 79
+    },
+    {
+      "epoch": 0.4199475065616798,
+      "grad_norm": 0.3853018091795304,
+      "learning_rate": 4.83118057351089e-06,
+      "loss": 0.9528,
+      "step": 80
+    },
+    {
+      "epoch": 0.4251968503937008,
+      "grad_norm": 0.41640976933906576,
+      "learning_rate": 4.822736515147748e-06,
+      "loss": 0.9281,
+      "step": 81
+    },
+    {
+      "epoch": 0.4304461942257218,
+      "grad_norm": 0.4341414623727107,
+      "learning_rate": 4.814094149453891e-06,
+      "loss": 0.983,
+      "step": 82
+    },
+    {
+      "epoch": 0.4356955380577428,
+      "grad_norm": 0.41893835966431175,
+      "learning_rate": 4.805254214285061e-06,
+      "loss": 0.9691,
+      "step": 83
+    },
+    {
+      "epoch": 0.4409448818897638,
+      "grad_norm": 0.3778203088676148,
+      "learning_rate": 4.796217464364808e-06,
+      "loss": 0.9386,
+      "step": 84
+    },
+    {
+      "epoch": 0.4461942257217848,
+      "grad_norm": 0.4049971835755209,
+      "learning_rate": 4.786984671220053e-06,
+      "loss": 0.9604,
+      "step": 85
+    },
+    {
+      "epoch": 0.45144356955380577,
+      "grad_norm": 0.39430021496671025,
+      "learning_rate": 4.7775566231152216e-06,
+      "loss": 0.9625,
+      "step": 86
+    },
+    {
+      "epoch": 0.4566929133858268,
+      "grad_norm": 0.4303135030837173,
+      "learning_rate": 4.767934124984941e-06,
+      "loss": 0.9421,
+      "step": 87
+    },
+    {
+      "epoch": 0.46194225721784776,
+      "grad_norm": 0.4037289734626591,
+      "learning_rate": 4.7581179983653224e-06,
+      "loss": 0.9395,
+      "step": 88
+    },
+    {
+      "epoch": 0.4671916010498688,
+      "grad_norm": 0.389977925556389,
+      "learning_rate": 4.7481090813238145e-06,
+      "loss": 0.9494,
+      "step": 89
+    },
+    {
+      "epoch": 0.47244094488188976,
+      "grad_norm": 0.4221791671439025,
+      "learning_rate": 4.737908228387656e-06,
+      "loss": 0.9236,
+      "step": 90
+    },
+    {
+      "epoch": 0.4776902887139108,
+      "grad_norm": 0.38571752654505004,
+      "learning_rate": 4.72751631047092e-06,
+      "loss": 0.9836,
+      "step": 91
+    },
+    {
+      "epoch": 0.48293963254593175,
+      "grad_norm": 0.3756683416227555,
+      "learning_rate": 4.716934214800155e-06,
+      "loss": 0.9847,
+      "step": 92
+    },
+    {
+      "epoch": 0.4881889763779528,
+      "grad_norm": 0.43805821160946656,
+      "learning_rate": 4.70616284483864e-06,
+      "loss": 0.9759,
+      "step": 93
+    },
+    {
+      "epoch": 0.49343832020997375,
+      "grad_norm": 0.42186430904629324,
+      "learning_rate": 4.695203120209245e-06,
+      "loss": 0.9381,
+      "step": 94
+    },
+    {
+      "epoch": 0.49868766404199477,
+      "grad_norm": 0.3816672833085737,
+      "learning_rate": 4.684055976615924e-06,
+      "loss": 0.9381,
+      "step": 95
+    },
+    {
+      "epoch": 0.5039370078740157,
+      "grad_norm": 0.4151475033902527,
+      "learning_rate": 4.672722365763821e-06,
+      "loss": 0.9449,
+      "step": 96
+    },
+    {
+      "epoch": 0.5091863517060368,
+      "grad_norm": 0.39782263887391656,
+      "learning_rate": 4.66120325527802e-06,
+      "loss": 0.9617,
+      "step": 97
+    },
+    {
+      "epoch": 0.5144356955380578,
+      "grad_norm": 0.40910931809868095,
+      "learning_rate": 4.649499628620931e-06,
+      "loss": 0.9334,
+      "step": 98
+    },
+    {
+      "epoch": 0.5196850393700787,
+      "grad_norm": 0.3910070221772484,
+      "learning_rate": 4.637612485008328e-06,
+      "loss": 0.9344,
+      "step": 99
+    },
+    {
+      "epoch": 0.5249343832020997,
+      "grad_norm": 0.3983780342477285,
+      "learning_rate": 4.625542839324036e-06,
+      "loss": 0.9201,
+      "step": 100
+    },
+    {
+      "epoch": 0.5301837270341208,
+      "grad_norm": 0.41440216503056093,
+      "learning_rate": 4.613291722033285e-06,
+      "loss": 0.9854,
+      "step": 101
+    },
+    {
+      "epoch": 0.5354330708661418,
+      "grad_norm": 0.399617487921351,
+      "learning_rate": 4.600860179094732e-06,
+      "loss": 0.9317,
+      "step": 102
+    },
+    {
+      "epoch": 0.5406824146981627,
+      "grad_norm": 0.38269422826680943,
+      "learning_rate": 4.588249271871164e-06,
+      "loss": 0.9026,
+      "step": 103
+    },
+    {
+      "epoch": 0.5459317585301837,
+      "grad_norm": 0.4161308186315819,
+      "learning_rate": 4.575460077038877e-06,
+      "loss": 0.9402,
+      "step": 104
+    },
+    {
+      "epoch": 0.5511811023622047,
+      "grad_norm": 0.38768080702286994,
+      "learning_rate": 4.562493686495756e-06,
+      "loss": 0.9276,
+      "step": 105
+    },
+    {
+      "epoch": 0.5564304461942258,
+      "grad_norm": 0.4454064598197059,
+      "learning_rate": 4.5493512072680535e-06,
+      "loss": 0.9452,
+      "step": 106
+    },
+    {
+      "epoch": 0.5616797900262467,
+      "grad_norm": 0.3695333796660741,
+      "learning_rate": 4.536033761415871e-06,
+      "loss": 0.9493,
+      "step": 107
+    },
+    {
+      "epoch": 0.5669291338582677,
+      "grad_norm": 0.4111826761670116,
+      "learning_rate": 4.522542485937369e-06,
+      "loss": 0.9109,
+      "step": 108
+    },
+    {
+      "epoch": 0.5721784776902887,
+      "grad_norm": 0.384299356893817,
+      "learning_rate": 4.508878532671684e-06,
+      "loss": 0.938,
+      "step": 109
+    },
+    {
+      "epoch": 0.5774278215223098,
+      "grad_norm": 0.39579785919606036,
+      "learning_rate": 4.4950430682005995e-06,
+      "loss": 0.924,
+      "step": 110
+    },
+    {
+      "epoch": 0.5826771653543307,
+      "grad_norm": 0.40856237862457445,
+      "learning_rate": 4.481037273748935e-06,
+      "loss": 0.9092,
+      "step": 111
+    },
+    {
+      "epoch": 0.5879265091863517,
+      "grad_norm": 0.38651275760437465,
+      "learning_rate": 4.4668623450837085e-06,
+      "loss": 0.9311,
+      "step": 112
+    },
+    {
+      "epoch": 0.5931758530183727,
+      "grad_norm": 0.378464413540759,
+      "learning_rate": 4.452519492412039e-06,
+      "loss": 0.9255,
+      "step": 113
+    },
+    {
+      "epoch": 0.5984251968503937,
+      "grad_norm": 0.3779607392785478,
+      "learning_rate": 4.438009940277825e-06,
+      "loss": 0.9024,
+      "step": 114
+    },
+    {
+      "epoch": 0.6036745406824147,
+      "grad_norm": 0.410919812420786,
+      "learning_rate": 4.423334927457198e-06,
+      "loss": 0.9105,
+      "step": 115
+    },
+    {
+      "epoch": 0.6089238845144357,
+      "grad_norm": 0.41911457133998964,
+      "learning_rate": 4.408495706852758e-06,
+      "loss": 0.9483,
+      "step": 116
+    },
+    {
+      "epoch": 0.6141732283464567,
+      "grad_norm": 0.39458100856937656,
+      "learning_rate": 4.393493545386607e-06,
+      "loss": 0.9388,
+      "step": 117
+    },
+    {
+      "epoch": 0.6194225721784777,
+      "grad_norm": 0.40200990450921853,
+      "learning_rate": 4.378329723892184e-06,
+      "loss": 0.927,
+      "step": 118
+    },
+    {
+      "epoch": 0.6246719160104987,
+      "grad_norm": 0.42346347209583185,
+      "learning_rate": 4.3630055370049065e-06,
+      "loss": 0.9439,
+      "step": 119
+    },
+    {
+      "epoch": 0.6299212598425197,
+      "grad_norm": 0.40678132773256936,
+      "learning_rate": 4.3475222930516484e-06,
+      "loss": 0.994,
+      "step": 120
+    },
+    {
+      "epoch": 0.6351706036745407,
+      "grad_norm": 0.4215857320252809,
+      "learning_rate": 4.3318813139390295e-06,
+      "loss": 0.8946,
+      "step": 121
+    },
+    {
+      "epoch": 0.6404199475065617,
+      "grad_norm": 0.39311078218513845,
+      "learning_rate": 4.316083935040561e-06,
+      "loss": 0.9129,
+      "step": 122
+    },
+    {
+      "epoch": 0.6456692913385826,
+      "grad_norm": 0.38855202294392815,
+      "learning_rate": 4.300131505082637e-06,
+      "loss": 0.9229,
+      "step": 123
+    },
+    {
+      "epoch": 0.6509186351706037,
+      "grad_norm": 0.437029643792513,
+      "learning_rate": 4.284025386029381e-06,
+      "loss": 0.9151,
+      "step": 124
+    },
+    {
+      "epoch": 0.6561679790026247,
+      "grad_norm": 0.41975844377898647,
+      "learning_rate": 4.267766952966369e-06,
+      "loss": 0.9224,
+      "step": 125
+    },
+    {
+      "epoch": 0.6614173228346457,
+      "grad_norm": 0.39675424131353054,
+      "learning_rate": 4.251357593983228e-06,
+      "loss": 0.9371,
+      "step": 126
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.3980469727389334,
+      "learning_rate": 4.234798710055124e-06,
+      "loss": 0.928,
+      "step": 127
+    },
+    {
+      "epoch": 0.6719160104986877,
+      "grad_norm": 0.3868709015712275,
+      "learning_rate": 4.218091714923157e-06,
+      "loss": 0.925,
+      "step": 128
+    },
+    {
+      "epoch": 0.6771653543307087,
+      "grad_norm": 0.4150283358784954,
+      "learning_rate": 4.2012380349736544e-06,
+      "loss": 0.9632,
+      "step": 129
+    },
+    {
+      "epoch": 0.6824146981627297,
+      "grad_norm": 0.3965769285465639,
+      "learning_rate": 4.184239109116393e-06,
+      "loss": 0.916,
+      "step": 130
+    },
+    {
+      "epoch": 0.6876640419947506,
+      "grad_norm": 0.4040770175389451,
+      "learning_rate": 4.167096388661754e-06,
+      "loss": 0.9013,
+      "step": 131
+    },
+    {
+      "epoch": 0.6929133858267716,
+      "grad_norm": 0.4065859256340288,
+      "learning_rate": 4.149811337196808e-06,
+      "loss": 0.9115,
+      "step": 132
+    },
+    {
+      "epoch": 0.6981627296587927,
+      "grad_norm": 0.39095752489168806,
+      "learning_rate": 4.132385430460361e-06,
+      "loss": 0.9401,
+      "step": 133
+    },
+    {
+      "epoch": 0.7034120734908137,
+      "grad_norm": 0.37457240781741025,
+      "learning_rate": 4.114820156216969e-06,
+      "loss": 0.9178,
+      "step": 134
+    },
+    {
+      "epoch": 0.7086614173228346,
+      "grad_norm": 0.399797934048477,
+      "learning_rate": 4.097117014129903e-06,
+      "loss": 0.9579,
+      "step": 135
+    },
+    {
+      "epoch": 0.7139107611548556,
+      "grad_norm": 0.4477976289426918,
+      "learning_rate": 4.079277515633127e-06,
+      "loss": 0.9448,
+      "step": 136
+    },
+    {
+      "epoch": 0.7191601049868767,
+      "grad_norm": 0.38322925064614577,
+      "learning_rate": 4.061303183802248e-06,
+      "loss": 0.9192,
+      "step": 137
+    },
+    {
+      "epoch": 0.7244094488188977,
+      "grad_norm": 0.40594305556063015,
+      "learning_rate": 4.043195553224482e-06,
+      "loss": 0.9254,
+      "step": 138
+    },
+    {
+      "epoch": 0.7296587926509186,
+      "grad_norm": 0.36120175115175723,
+      "learning_rate": 4.024956169867642e-06,
+      "loss": 0.9591,
+      "step": 139
+    },
+    {
+      "epoch": 0.7349081364829396,
+      "grad_norm": 0.38003421628995726,
+      "learning_rate": 4.006586590948141e-06,
+      "loss": 0.9154,
+      "step": 140
+    },
+    {
+      "epoch": 0.7401574803149606,
+      "grad_norm": 0.3779196146369032,
+      "learning_rate": 3.9880883847980475e-06,
+      "loss": 0.9412,
+      "step": 141
+    },
+    {
+      "epoch": 0.7454068241469817,
+      "grad_norm": 0.3908051818324559,
+      "learning_rate": 3.969463130731183e-06,
+      "loss": 0.9037,
+      "step": 142
+    },
+    {
+      "epoch": 0.7506561679790026,
+      "grad_norm": 0.4043989800878751,
+      "learning_rate": 3.95071241890829e-06,
+      "loss": 0.9032,
+      "step": 143
+    },
+    {
+      "epoch": 0.7559055118110236,
+      "grad_norm": 0.396255171092691,
+      "learning_rate": 3.9318378502012636e-06,
+      "loss": 0.9288,
+      "step": 144
+    },
+    {
+      "epoch": 0.7611548556430446,
+      "grad_norm": 0.40757200005039956,
+      "learning_rate": 3.91284103605648e-06,
+      "loss": 0.9179,
+      "step": 145
+    },
+    {
+      "epoch": 0.7664041994750657,
+      "grad_norm": 0.3862826523020415,
+      "learning_rate": 3.893723598357214e-06,
+      "loss": 0.8894,
+      "step": 146
+    },
+    {
+      "epoch": 0.7716535433070866,
+      "grad_norm": 0.3770664118848218,
+      "learning_rate": 3.874487169285168e-06,
+      "loss": 0.8898,
+      "step": 147
+    },
+    {
+      "epoch": 0.7769028871391076,
+      "grad_norm": 0.3854873052422692,
+      "learning_rate": 3.855133391181124e-06,
+      "loss": 0.9135,
+      "step": 148
+    },
+    {
+      "epoch": 0.7821522309711286,
+      "grad_norm": 0.38997928953459443,
+      "learning_rate": 3.835663916404721e-06,
+      "loss": 0.8843,
+      "step": 149
+    },
+    {
+      "epoch": 0.7874015748031497,
+      "grad_norm": 0.45712052623884775,
+      "learning_rate": 3.81608040719339e-06,
+      "loss": 0.9694,
+      "step": 150
+    },
+    {
+      "epoch": 0.7926509186351706,
+      "grad_norm": 0.41182630339038373,
+      "learning_rate": 3.7963845355204303e-06,
+      "loss": 0.9194,
+      "step": 151
+    },
+    {
+      "epoch": 0.7979002624671916,
+      "grad_norm": 0.4065794661619882,
+      "learning_rate": 3.7765779829522674e-06,
+      "loss": 0.9278,
+      "step": 152
+    },
+    {
+      "epoch": 0.8031496062992126,
+      "grad_norm": 0.38122636394786497,
+      "learning_rate": 3.7566624405048847e-06,
+      "loss": 0.892,
+      "step": 153
+    },
+    {
+      "epoch": 0.8083989501312336,
+      "grad_norm": 0.4026809828765533,
+      "learning_rate": 3.736639608499448e-06,
+      "loss": 0.9246,
+      "step": 154
+    },
+    {
+      "epoch": 0.8136482939632546,
+      "grad_norm": 0.3884561226623423,
+      "learning_rate": 3.7165111964171407e-06,
+      "loss": 0.9438,
+      "step": 155
+    },
+    {
+      "epoch": 0.8188976377952756,
+      "grad_norm": 0.3844815855908215,
+      "learning_rate": 3.6962789227532165e-06,
+      "loss": 0.9316,
+      "step": 156
+    },
+    {
+      "epoch": 0.8241469816272966,
+      "grad_norm": 0.3715029261772477,
+      "learning_rate": 3.675944514870274e-06,
+      "loss": 0.924,
+      "step": 157
+    },
+    {
+      "epoch": 0.8293963254593176,
+      "grad_norm": 0.4062567635907239,
+      "learning_rate": 3.6555097088507837e-06,
+      "loss": 0.9616,
+      "step": 158
+    },
+    {
+      "epoch": 0.8346456692913385,
+      "grad_norm": 0.40370287012056855,
+      "learning_rate": 3.634976249348867e-06,
+      "loss": 0.9526,
+      "step": 159
+    },
+    {
+      "epoch": 0.8398950131233596,
+      "grad_norm": 0.3899308479039462,
+      "learning_rate": 3.6143458894413463e-06,
+      "loss": 0.9215,
+      "step": 160
+    },
+    {
+      "epoch": 0.8451443569553806,
+      "grad_norm": 0.3813874729696962,
+      "learning_rate": 3.5936203904780665e-06,
+      "loss": 0.953,
+      "step": 161
+    },
+    {
+      "epoch": 0.8503937007874016,
+      "grad_norm": 0.38224710953621555,
+      "learning_rate": 3.5728015219315226e-06,
+      "loss": 0.894,
+      "step": 162
+    },
+    {
+      "epoch": 0.8556430446194225,
+      "grad_norm": 0.4038493099805114,
+      "learning_rate": 3.5518910612457885e-06,
+      "loss": 0.9614,
+      "step": 163
+    },
+    {
+      "epoch": 0.8608923884514436,
+      "grad_norm": 0.3976228258032158,
+      "learning_rate": 3.530890793684759e-06,
+      "loss": 0.9364,
+      "step": 164
+    },
+    {
+      "epoch": 0.8661417322834646,
+      "grad_norm": 0.4141980981190029,
+      "learning_rate": 3.5098025121797375e-06,
+      "loss": 0.9316,
+      "step": 165
+    },
+    {
+      "epoch": 0.8713910761154856,
+      "grad_norm": 0.4156903411116242,
+      "learning_rate": 3.4886280171763563e-06,
+      "loss": 0.923,
+      "step": 166
+    },
+    {
+      "epoch": 0.8766404199475065,
+      "grad_norm": 0.38250613966133334,
+      "learning_rate": 3.467369116480864e-06,
+      "loss": 0.9153,
+      "step": 167
+    },
+    {
+      "epoch": 0.8818897637795275,
+      "grad_norm": 0.4070344049084728,
+      "learning_rate": 3.446027625105776e-06,
+      "loss": 0.9347,
+      "step": 168
+    },
+    {
+      "epoch": 0.8871391076115486,
+      "grad_norm": 0.3844877252304378,
+      "learning_rate": 3.424605365114923e-06,
+      "loss": 0.9214,
+      "step": 169
+    },
+    {
+      "epoch": 0.8923884514435696,
+      "grad_norm": 0.37585915847896717,
+      "learning_rate": 3.403104165467883e-06,
+      "loss": 0.9133,
+      "step": 170
+    },
+    {
+      "epoch": 0.8976377952755905,
+      "grad_norm": 0.6300998021233689,
+      "learning_rate": 3.3815258618638316e-06,
+      "loss": 0.9395,
+      "step": 171
+    },
+    {
+      "epoch": 0.9028871391076115,
+      "grad_norm": 0.3994171061317929,
+      "learning_rate": 3.359872296584821e-06,
+      "loss": 0.917,
+      "step": 172
+    },
+    {
+      "epoch": 0.9081364829396326,
+      "grad_norm": 0.3806190962949758,
+      "learning_rate": 3.338145318338485e-06,
+      "loss": 0.9408,
+      "step": 173
+    },
+    {
+      "epoch": 0.9133858267716536,
+      "grad_norm": 0.36777162273585867,
+      "learning_rate": 3.3163467821002082e-06,
+      "loss": 0.9346,
+      "step": 174
+    },
+    {
+      "epoch": 0.9186351706036745,
+      "grad_norm": 0.3927948155295108,
+      "learning_rate": 3.2944785489547544e-06,
+      "loss": 0.9121,
+      "step": 175
+    },
+    {
+      "epoch": 0.9238845144356955,
+      "grad_norm": 0.3958182964876464,
+      "learning_rate": 3.272542485937369e-06,
+      "loss": 0.9318,
+      "step": 176
+    },
+    {
+      "epoch": 0.9291338582677166,
+      "grad_norm": 0.39754399123912254,
+      "learning_rate": 3.250540465874382e-06,
+      "loss": 0.9244,
+      "step": 177
+    },
+    {
+      "epoch": 0.9343832020997376,
+      "grad_norm": 0.38861358420887904,
+      "learning_rate": 3.228474367223312e-06,
+      "loss": 0.9051,
+      "step": 178
+    },
+    {
+      "epoch": 0.9396325459317585,
+      "grad_norm": 0.3926071787199394,
+      "learning_rate": 3.206346073912488e-06,
+      "loss": 0.9409,
+      "step": 179
+    },
+    {
+      "epoch": 0.9448818897637795,
+      "grad_norm": 0.4008739814562732,
+      "learning_rate": 3.184157475180208e-06,
+      "loss": 0.9222,
+      "step": 180
+    },
+    {
+      "epoch": 0.9501312335958005,
+      "grad_norm": 0.3665607732753151,
+      "learning_rate": 3.1619104654134397e-06,
+      "loss": 0.913,
+      "step": 181
+    },
+    {
+      "epoch": 0.9553805774278216,
+      "grad_norm": 0.38983081632202093,
+      "learning_rate": 3.1396069439860894e-06,
+      "loss": 0.9297,
+      "step": 182
+    },
+    {
+      "epoch": 0.9606299212598425,
+      "grad_norm": 0.3645800135143814,
+      "learning_rate": 3.117248815096833e-06,
+      "loss": 0.8883,
+      "step": 183
+    },
+    {
+      "epoch": 0.9658792650918635,
+      "grad_norm": 0.3720108758786826,
+      "learning_rate": 3.094837987606547e-06,
+      "loss": 0.9204,
+      "step": 184
+    },
+    {
+      "epoch": 0.9711286089238845,
+      "grad_norm": 0.3818234182095755,
+      "learning_rate": 3.0723763748753354e-06,
+      "loss": 0.8814,
+      "step": 185
+    },
+    {
+      "epoch": 0.9763779527559056,
+      "grad_norm": 0.3847151463563777,
+      "learning_rate": 3.049865894599172e-06,
+      "loss": 0.9133,
+      "step": 186
+    },
+    {
+      "epoch": 0.9816272965879265,
+      "grad_norm": 0.37971775342950864,
+      "learning_rate": 3.027308468646175e-06,
+      "loss": 0.8906,
+      "step": 187
+    },
+    {
+      "epoch": 0.9868766404199475,
+      "grad_norm": 0.41572124992250203,
+      "learning_rate": 3.0047060228925256e-06,
+      "loss": 0.9672,
+      "step": 188
+    },
+    {
+      "epoch": 0.9921259842519685,
+      "grad_norm": 0.40219608509658256,
+      "learning_rate": 2.9820604870580426e-06,
+      "loss": 0.9011,
+      "step": 189
+    },
+    {
+      "epoch": 0.9973753280839895,
+      "grad_norm": 0.397234257384178,
+      "learning_rate": 2.9593737945414264e-06,
+      "loss": 0.9174,
+      "step": 190
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.397234257384178,
+      "learning_rate": 2.9366478822551973e-06,
+      "loss": 0.9204,
+      "step": 191
+    },
+    {
+      "epoch": 1.005249343832021,
+      "grad_norm": 0.6279719871788059,
+      "learning_rate": 2.913884690460325e-06,
+      "loss": 0.8376,
+      "step": 192
+    },
+    {
+      "epoch": 1.010498687664042,
+      "grad_norm": 0.41242697308859033,
+      "learning_rate": 2.8910861626005774e-06,
+      "loss": 0.8369,
+      "step": 193
+    },
+    {
+      "epoch": 1.015748031496063,
+      "grad_norm": 0.38981163388237194,
+      "learning_rate": 2.8682542451365943e-06,
+      "loss": 0.8493,
+      "step": 194
+    },
+    {
+      "epoch": 1.020997375328084,
+      "grad_norm": 0.42318354204075453,
+      "learning_rate": 2.845390887379706e-06,
+      "loss": 0.8618,
+      "step": 195
+    },
+    {
+      "epoch": 1.026246719160105,
+      "grad_norm": 0.5052809667467608,
+      "learning_rate": 2.822498041325509e-06,
+      "loss": 0.8644,
+      "step": 196
+    },
+    {
+      "epoch": 1.031496062992126,
+      "grad_norm": 0.36960238923766053,
+      "learning_rate": 2.7995776614872083e-06,
+      "loss": 0.8484,
+      "step": 197
+    },
+    {
+      "epoch": 1.036745406824147,
+      "grad_norm": 0.3606472920225704,
+      "learning_rate": 2.776631704728752e-06,
+      "loss": 0.8413,
+      "step": 198
+    },
+    {
+      "epoch": 1.041994750656168,
+      "grad_norm": 0.38778609923815943,
+      "learning_rate": 2.753662130097758e-06,
+      "loss": 0.8266,
+      "step": 199
+    },
+    {
+      "epoch": 1.047244094488189,
+      "grad_norm": 0.3636856280047818,
+      "learning_rate": 2.730670898658255e-06,
+      "loss": 0.8285,
+      "step": 200
+    },
+    {
+      "epoch": 1.05249343832021,
+      "grad_norm": 0.3886565787437705,
+      "learning_rate": 2.70765997332326e-06,
+      "loss": 0.8628,
+      "step": 201
+    },
+    {
+      "epoch": 1.057742782152231,
+      "grad_norm": 0.41378173196429036,
+      "learning_rate": 2.684631318687185e-06,
+      "loss": 0.8549,
+      "step": 202
+    },
+    {
+      "epoch": 1.0629921259842519,
+      "grad_norm": 0.3657527566283362,
+      "learning_rate": 2.661586900858111e-06,
+      "loss": 0.8472,
+      "step": 203
+    },
+    {
+      "epoch": 1.068241469816273,
+      "grad_norm": 0.3666340150026852,
+      "learning_rate": 2.638528687289925e-06,
+      "loss": 0.8331,
+      "step": 204
+    },
+    {
+      "epoch": 1.073490813648294,
+      "grad_norm": 0.3661321106045701,
+      "learning_rate": 2.6154586466143495e-06,
+      "loss": 0.8706,
+      "step": 205
+    },
+    {
+      "epoch": 1.078740157480315,
+      "grad_norm": 0.3796204270047528,
+      "learning_rate": 2.592378748472863e-06,
+      "loss": 0.8329,
+      "step": 206
+    },
+    {
+      "epoch": 1.083989501312336,
+      "grad_norm": 0.4003268539729557,
+      "learning_rate": 2.5692909633485414e-06,
+      "loss": 0.8762,
+      "step": 207
+    },
+    {
+      "epoch": 1.0892388451443569,
+      "grad_norm": 0.3816091507612548,
+      "learning_rate": 2.546197262397825e-06,
+      "loss": 0.8499,
+      "step": 208
+    },
+    {
+      "epoch": 1.094488188976378,
+      "grad_norm": 0.4135433735229758,
+      "learning_rate": 2.5230996172822274e-06,
+      "loss": 0.8191,
+      "step": 209
+    },
+    {
+      "epoch": 1.099737532808399,
+      "grad_norm": 0.38519176727175336,
+      "learning_rate": 2.5e-06,
+      "loss": 0.8164,
+      "step": 210
+    },
+    {
+      "epoch": 1.10498687664042,
+      "grad_norm": 0.397224934471037,
+      "learning_rate": 2.4769003827177735e-06,
+      "loss": 0.8373,
+      "step": 211
+    },
+    {
+      "epoch": 1.110236220472441,
+      "grad_norm": 0.37696141546345585,
+      "learning_rate": 2.453802737602176e-06,
+      "loss": 0.8575,
+      "step": 212
+    },
+    {
+      "epoch": 1.1154855643044619,
+      "grad_norm": 0.38562550508165394,
+      "learning_rate": 2.4307090366514594e-06,
+      "loss": 0.8372,
+      "step": 213
+    },
+    {
+      "epoch": 1.120734908136483,
+      "grad_norm": 0.392610655726213,
+      "learning_rate": 2.4076212515271384e-06,
+      "loss": 0.8561,
+      "step": 214
+    },
+    {
+      "epoch": 1.125984251968504,
+      "grad_norm": 0.3752434088251031,
+      "learning_rate": 2.3845413533856517e-06,
+      "loss": 0.8539,
+      "step": 215
+    },
+    {
+      "epoch": 1.1312335958005248,
+      "grad_norm": 0.3769283394800778,
+      "learning_rate": 2.3614713127100752e-06,
+      "loss": 0.8134,
+      "step": 216
+    },
+    {
+      "epoch": 1.136482939632546,
+      "grad_norm": 0.39533971296250825,
+      "learning_rate": 2.3384130991418896e-06,
+      "loss": 0.8608,
+      "step": 217
+    },
+    {
+      "epoch": 1.141732283464567,
+      "grad_norm": 0.5791866799791944,
+      "learning_rate": 2.3153686813128153e-06,
+      "loss": 0.8334,
+      "step": 218
+    },
+    {
+      "epoch": 1.1469816272965878,
+      "grad_norm": 0.3810153033409976,
+      "learning_rate": 2.2923400266767406e-06,
+      "loss": 0.8472,
+      "step": 219
+    },
+    {
+      "epoch": 1.152230971128609,
+      "grad_norm": 0.3728143509479016,
+      "learning_rate": 2.269329101341745e-06,
+      "loss": 0.8376,
+      "step": 220
+    },
+    {
+      "epoch": 1.1574803149606299,
+      "grad_norm": 0.40305096887095054,
+      "learning_rate": 2.246337869902243e-06,
+      "loss": 0.8275,
+      "step": 221
+    },
+    {
+      "epoch": 1.162729658792651,
+      "grad_norm": 0.3692851271916895,
+      "learning_rate": 2.2233682952712484e-06,
+      "loss": 0.8413,
+      "step": 222
+    },
+    {
+      "epoch": 1.167979002624672,
+      "grad_norm": 0.3963527736664121,
+      "learning_rate": 2.2004223385127925e-06,
+      "loss": 0.8295,
+      "step": 223
+    },
+    {
+      "epoch": 1.1732283464566928,
+      "grad_norm": 0.3587695306587395,
+      "learning_rate": 2.1775019586744924e-06,
+      "loss": 0.8547,
+      "step": 224
+    },
+    {
+      "epoch": 1.178477690288714,
+      "grad_norm": 0.3614709935886563,
+      "learning_rate": 2.1546091126202955e-06,
+      "loss": 0.813,
+      "step": 225
+    },
+    {
+      "epoch": 1.1837270341207349,
+      "grad_norm": 0.4140010983968987,
+      "learning_rate": 2.131745754863406e-06,
+      "loss": 0.8398,
+      "step": 226
+    },
+    {
+      "epoch": 1.188976377952756,
+      "grad_norm": 0.45996885685295197,
+      "learning_rate": 2.1089138373994226e-06,
+      "loss": 0.8518,
+      "step": 227
+    },
+    {
+      "epoch": 1.194225721784777,
+      "grad_norm": 0.36254259181474985,
+      "learning_rate": 2.086115309539675e-06,
+      "loss": 0.8647,
+      "step": 228
+    },
+    {
+      "epoch": 1.1994750656167978,
+      "grad_norm": 0.4122269000782737,
+      "learning_rate": 2.063352117744803e-06,
+      "loss": 0.8881,
+      "step": 229
+    },
+    {
+      "epoch": 1.204724409448819,
+      "grad_norm": 0.40615776242804974,
+      "learning_rate": 2.040626205458574e-06,
+      "loss": 0.8328,
+      "step": 230
+    },
+    {
+      "epoch": 1.20997375328084,
+      "grad_norm": 0.41826600558691523,
+      "learning_rate": 2.017939512941958e-06,
+      "loss": 0.8281,
+      "step": 231
+    },
+    {
+      "epoch": 1.2152230971128608,
+      "grad_norm": 0.4153271729906844,
+      "learning_rate": 1.995293977107475e-06,
+      "loss": 0.8693,
+      "step": 232
+    },
+    {
+      "epoch": 1.220472440944882,
+      "grad_norm": 0.3654074693662248,
+      "learning_rate": 1.972691531353826e-06,
+      "loss": 0.821,
+      "step": 233
+    },
+    {
+      "epoch": 1.2257217847769029,
+      "grad_norm": 0.3981178374801672,
+      "learning_rate": 1.9501341054008292e-06,
+      "loss": 0.8962,
+      "step": 234
+    },
+    {
+      "epoch": 1.2309711286089238,
+      "grad_norm": 0.4049324135862524,
+      "learning_rate": 1.9276236251246655e-06,
+      "loss": 0.7905,
+      "step": 235
+    },
+    {
+      "epoch": 1.236220472440945,
+      "grad_norm": 0.33631867862267323,
+      "learning_rate": 1.9051620123934538e-06,
+      "loss": 0.8284,
+      "step": 236
+    },
+    {
+      "epoch": 1.2414698162729658,
+      "grad_norm": 0.37393386640045784,
+      "learning_rate": 1.882751184903167e-06,
+      "loss": 0.8405,
+      "step": 237
+    },
+    {
+      "epoch": 1.246719160104987,
+      "grad_norm": 0.3462880519364805,
+      "learning_rate": 1.860393056013911e-06,
+      "loss": 0.7939,
+      "step": 238
+    },
+    {
+      "epoch": 1.2519685039370079,
+      "grad_norm": 0.3848360075044728,
+      "learning_rate": 1.8380895345865603e-06,
+      "loss": 0.8375,
+      "step": 239
+    },
+    {
+      "epoch": 1.257217847769029,
+      "grad_norm": 0.3617106402936481,
+      "learning_rate": 1.8158425248197931e-06,
+      "loss": 0.8162,
+      "step": 240
+    },
+    {
+      "epoch": 1.26246719160105,
+      "grad_norm": 0.34772753646733273,
+      "learning_rate": 1.7936539260875125e-06,
+      "loss": 0.8408,
+      "step": 241
+    },
+    {
+      "epoch": 1.2677165354330708,
+      "grad_norm": 0.36406532287683085,
+      "learning_rate": 1.7715256327766887e-06,
+      "loss": 0.8103,
+      "step": 242
+    },
+    {
+      "epoch": 1.272965879265092,
+      "grad_norm": 0.3765574533307364,
+      "learning_rate": 1.7494595341256185e-06,
+      "loss": 0.8461,
+      "step": 243
+    },
+    {
+      "epoch": 1.2782152230971129,
+      "grad_norm": 0.3859435308304487,
+      "learning_rate": 1.7274575140626318e-06,
+      "loss": 0.8761,
+      "step": 244
+    },
+    {
+      "epoch": 1.2834645669291338,
+      "grad_norm": 0.3936691957558663,
+      "learning_rate": 1.7055214510452462e-06,
+      "loss": 0.9159,
+      "step": 245
+    },
+    {
+      "epoch": 1.288713910761155,
+      "grad_norm": 0.5107365379023212,
+      "learning_rate": 1.6836532178997922e-06,
+      "loss": 0.8649,
+      "step": 246
+    },
+    {
+      "epoch": 1.2939632545931758,
+      "grad_norm": 0.388972979502565,
+      "learning_rate": 1.6618546816615162e-06,
+      "loss": 0.8734,
+      "step": 247
+    },
+    {
+      "epoch": 1.2992125984251968,
+      "grad_norm": 0.4410810635653612,
+      "learning_rate": 1.6401277034151798e-06,
+      "loss": 0.8405,
+      "step": 248
+    },
+    {
+      "epoch": 1.304461942257218,
+      "grad_norm": 0.38502463729196623,
+      "learning_rate": 1.6184741381361684e-06,
+      "loss": 0.8546,
+      "step": 249
+    },
+    {
+      "epoch": 1.3097112860892388,
+      "grad_norm": 0.38032181472927906,
+      "learning_rate": 1.5968958345321178e-06,
+      "loss": 0.8253,
+      "step": 250
+    },
+    {
+      "epoch": 1.3149606299212597,
+      "grad_norm": 0.36970392450888717,
+      "learning_rate": 1.5753946348850774e-06,
+      "loss": 0.8558,
+      "step": 251
+    },
+    {
+      "epoch": 1.3202099737532809,
+      "grad_norm": 0.3934484470589474,
+      "learning_rate": 1.5539723748942246e-06,
+      "loss": 0.8299,
+      "step": 252
+    },
+    {
+      "epoch": 1.3254593175853018,
+      "grad_norm": 0.3602834073019444,
+      "learning_rate": 1.5326308835191372e-06,
+      "loss": 0.8476,
+      "step": 253
+    },
+    {
+      "epoch": 1.330708661417323,
+      "grad_norm": 0.3482676328694225,
+      "learning_rate": 1.5113719828236439e-06,
+      "loss": 0.858,
+      "step": 254
+    },
+    {
+      "epoch": 1.3359580052493438,
+      "grad_norm": 0.44035367856347457,
+      "learning_rate": 1.490197487820263e-06,
+      "loss": 0.8351,
+      "step": 255
+    },
+    {
+      "epoch": 1.341207349081365,
+      "grad_norm": 0.371715281121202,
+      "learning_rate": 1.4691092063152417e-06,
+      "loss": 0.827,
+      "step": 256
+    },
+    {
+      "epoch": 1.3464566929133859,
+      "grad_norm": 0.35070038567136974,
+      "learning_rate": 1.4481089387542134e-06,
+      "loss": 0.8492,
+      "step": 257
+    },
+    {
+      "epoch": 1.3517060367454068,
+      "grad_norm": 0.3499117772233927,
+      "learning_rate": 1.4271984780684778e-06,
+      "loss": 0.8399,
+      "step": 258
+    },
+    {
+      "epoch": 1.356955380577428,
+      "grad_norm": 0.3811483044855955,
+      "learning_rate": 1.4063796095219345e-06,
+      "loss": 0.8688,
+      "step": 259
+    },
+    {
+      "epoch": 1.3622047244094488,
+      "grad_norm": 0.3714007196954483,
+      "learning_rate": 1.3856541105586545e-06,
+      "loss": 0.813,
+      "step": 260
+    },
+    {
+      "epoch": 1.3674540682414698,
+      "grad_norm": 0.3577252388966486,
+      "learning_rate": 1.3650237506511333e-06,
+      "loss": 0.8506,
+      "step": 261
+    },
+    {
+      "epoch": 1.372703412073491,
+      "grad_norm": 0.3670679328461459,
+      "learning_rate": 1.3444902911492174e-06,
+      "loss": 0.8267,
+      "step": 262
+    },
+    {
+      "epoch": 1.3779527559055118,
+      "grad_norm": 0.3521888751612076,
+      "learning_rate": 1.324055485129727e-06,
+      "loss": 0.8079,
+      "step": 263
+    },
+    {
+      "epoch": 1.3832020997375327,
+      "grad_norm": 0.3751986141403746,
+      "learning_rate": 1.303721077246784e-06,
+      "loss": 0.8491,
+      "step": 264
+    },
+    {
+      "epoch": 1.3884514435695539,
+      "grad_norm": 0.3589588384505892,
+      "learning_rate": 1.2834888035828597e-06,
+      "loss": 0.8768,
+      "step": 265
+    },
+    {
+      "epoch": 1.3937007874015748,
+      "grad_norm": 0.372280310354174,
+      "learning_rate": 1.2633603915005535e-06,
+      "loss": 0.8335,
+      "step": 266
+    },
+    {
+      "epoch": 1.3989501312335957,
+      "grad_norm": 0.40534475247811924,
+      "learning_rate": 1.2433375594951166e-06,
+      "loss": 0.8719,
+      "step": 267
+    },
+    {
+      "epoch": 1.4041994750656168,
+      "grad_norm": 0.40997625659212245,
+      "learning_rate": 1.2234220170477332e-06,
+      "loss": 0.8269,
+      "step": 268
+    },
+    {
+      "epoch": 1.4094488188976377,
+      "grad_norm": 0.3668404917652971,
+      "learning_rate": 1.2036154644795697e-06,
+      "loss": 0.7913,
+      "step": 269
+    },
+    {
+      "epoch": 1.4146981627296589,
+      "grad_norm": 0.3641040286199539,
+      "learning_rate": 1.1839195928066101e-06,
+      "loss": 0.8469,
+      "step": 270
+    },
+    {
+      "epoch": 1.4199475065616798,
+      "grad_norm": 0.36110674154533795,
+      "learning_rate": 1.164336083595279e-06,
+      "loss": 0.8698,
+      "step": 271
+    },
+    {
+      "epoch": 1.425196850393701,
+      "grad_norm": 0.376323216718325,
+      "learning_rate": 1.1448666088188766e-06,
+      "loss": 0.8355,
+      "step": 272
+    },
+    {
+      "epoch": 1.4304461942257218,
+      "grad_norm": 0.37039873117700245,
+      "learning_rate": 1.1255128307148319e-06,
+      "loss": 0.8219,
+      "step": 273
+    },
+    {
+      "epoch": 1.4356955380577427,
+      "grad_norm": 0.3612230977145218,
+      "learning_rate": 1.1062764016427864e-06,
+      "loss": 0.8568,
+      "step": 274
+    },
+    {
+      "epoch": 1.4409448818897639,
+      "grad_norm": 0.3962177187198973,
+      "learning_rate": 1.0871589639435204e-06,
+      "loss": 0.8115,
+      "step": 275
+    },
+    {
+      "epoch": 1.4461942257217848,
+      "grad_norm": 0.38303094180829605,
+      "learning_rate": 1.068162149798737e-06,
+      "loss": 0.818,
+      "step": 276
+    },
+    {
+      "epoch": 1.4514435695538057,
+      "grad_norm": 0.364448052838668,
+      "learning_rate": 1.049287581091711e-06,
+      "loss": 0.803,
+      "step": 277
+    },
+    {
+      "epoch": 1.4566929133858268,
+      "grad_norm": 0.37746023670056045,
+      "learning_rate": 1.0305368692688175e-06,
+      "loss": 0.8495,
+      "step": 278
+    },
+    {
+      "epoch": 1.4619422572178478,
+      "grad_norm": 0.40984945067844814,
+      "learning_rate": 1.0119116152019535e-06,
+      "loss": 0.8631,
+      "step": 279
+    },
+    {
+      "epoch": 1.4671916010498687,
+      "grad_norm": 0.3840327770579671,
+      "learning_rate": 9.934134090518593e-07,
+      "loss": 0.8318,
+      "step": 280
+    },
+    {
+      "epoch": 1.4724409448818898,
+      "grad_norm": 0.3575557636540703,
+      "learning_rate": 9.750438301323584e-07,
+      "loss": 0.8244,
+      "step": 281
+    },
+    {
+      "epoch": 1.4776902887139107,
+      "grad_norm": 0.363642443404998,
+      "learning_rate": 9.56804446775518e-07,
+      "loss": 0.8109,
+      "step": 282
+    },
+    {
+      "epoch": 1.4829396325459316,
+      "grad_norm": 0.3410295671236108,
+      "learning_rate": 9.386968161977528e-07,
+      "loss": 0.8227,
+      "step": 283
+    },
+    {
+      "epoch": 1.4881889763779528,
+      "grad_norm": 0.3693264048471325,
+      "learning_rate": 9.207224843668733e-07,
+      "loss": 0.8349,
+      "step": 284
+    },
+    {
+      "epoch": 1.4934383202099737,
+      "grad_norm": 0.3912247837861982,
+      "learning_rate": 9.028829858700974e-07,
+      "loss": 0.8218,
+      "step": 285
+    },
+    {
+      "epoch": 1.4986876640419948,
+      "grad_norm": 0.3771536484006796,
+      "learning_rate": 8.851798437830323e-07,
+      "loss": 0.8328,
+      "step": 286
+    },
+    {
+      "epoch": 1.5039370078740157,
+      "grad_norm": 0.3956914054034924,
+      "learning_rate": 8.676145695396399e-07,
+      "loss": 0.8779,
+      "step": 287
+    },
+    {
+      "epoch": 1.5091863517060369,
+      "grad_norm": 0.34503047792761804,
+      "learning_rate": 8.501886628031941e-07,
+      "loss": 0.8265,
+      "step": 288
+    },
+    {
+      "epoch": 1.5144356955380578,
+      "grad_norm": 0.3888965654379534,
+      "learning_rate": 8.329036113382474e-07,
+      "loss": 0.8354,
+      "step": 289
+    },
+    {
+      "epoch": 1.5196850393700787,
+      "grad_norm": 0.38477861699293,
+      "learning_rate": 8.157608908836071e-07,
+      "loss": 0.8109,
+      "step": 290
+    },
+    {
+      "epoch": 1.5249343832020998,
+      "grad_norm": 0.378643274506092,
+      "learning_rate": 7.987619650263462e-07,
+      "loss": 0.8507,
+      "step": 291
+    },
+    {
+      "epoch": 1.5301837270341208,
+      "grad_norm": 0.3619947185774947,
+      "learning_rate": 7.819082850768433e-07,
+      "loss": 0.8321,
+      "step": 292
+    },
+    {
+      "epoch": 1.5354330708661417,
+      "grad_norm": 0.3793890733316615,
+      "learning_rate": 7.652012899448761e-07,
+      "loss": 0.846,
+      "step": 293
+    },
+    {
+      "epoch": 1.5406824146981628,
+      "grad_norm": 0.37104232865461917,
+      "learning_rate": 7.486424060167726e-07,
+      "loss": 0.8113,
+      "step": 294
+    },
+    {
+      "epoch": 1.5459317585301837,
+      "grad_norm": 2.1051715634012864,
+      "learning_rate": 7.322330470336314e-07,
+      "loss": 0.8174,
+      "step": 295
+    },
+    {
+      "epoch": 1.5511811023622046,
+      "grad_norm": 0.45390059444688674,
+      "learning_rate": 7.159746139706194e-07,
+      "loss": 0.8414,
+      "step": 296
+    },
+    {
+      "epoch": 1.5564304461942258,
+      "grad_norm": 0.39115421496994207,
+      "learning_rate": 6.99868494917364e-07,
+      "loss": 0.8085,
+      "step": 297
+    },
+    {
+      "epoch": 1.5616797900262467,
+      "grad_norm": 0.37250458373292433,
+      "learning_rate": 6.839160649594401e-07,
+      "loss": 0.7906,
+      "step": 298
+    },
+    {
+      "epoch": 1.5669291338582676,
+      "grad_norm": 0.3746838213451273,
+      "learning_rate": 6.68118686060972e-07,
+      "loss": 0.8314,
+      "step": 299
+    },
+    {
+      "epoch": 1.5721784776902887,
+      "grad_norm": 0.5365749531577879,
+      "learning_rate": 6.524777069483526e-07,
+      "loss": 0.8332,
+      "step": 300
+    },
+    {
+      "epoch": 1.5774278215223099,
+      "grad_norm": 0.3788491507367419,
+      "learning_rate": 6.369944629950933e-07,
+      "loss": 0.8611,
+      "step": 301
+    },
+    {
+      "epoch": 1.5826771653543306,
+      "grad_norm": 0.3690009715495826,
+      "learning_rate": 6.216702761078167e-07,
+      "loss": 0.8099,
+      "step": 302
+    },
+    {
+      "epoch": 1.5879265091863517,
+      "grad_norm": 0.34643553497540036,
+      "learning_rate": 6.06506454613393e-07,
+      "loss": 0.8255,
+      "step": 303
+    },
+    {
+      "epoch": 1.5931758530183728,
+      "grad_norm": 0.37384009617772274,
+      "learning_rate": 5.915042931472426e-07,
+      "loss": 0.8024,
+      "step": 304
+    },
+    {
+      "epoch": 1.5984251968503937,
+      "grad_norm": 0.34112911330910745,
+      "learning_rate": 5.766650725428027e-07,
+      "loss": 0.8172,
+      "step": 305
+    },
+    {
+      "epoch": 1.6036745406824147,
+      "grad_norm": 0.3675924179740712,
+      "learning_rate": 5.619900597221753e-07,
+      "loss": 0.8195,
+      "step": 306
+    },
+    {
+      "epoch": 1.6089238845144358,
+      "grad_norm": 0.36296097997746496,
+      "learning_rate": 5.474805075879616e-07,
+      "loss": 0.808,
+      "step": 307
+    },
+    {
+      "epoch": 1.6141732283464567,
+      "grad_norm": 0.37901033900237063,
+      "learning_rate": 5.33137654916292e-07,
+      "loss": 0.8068,
+      "step": 308
+    },
+    {
+      "epoch": 1.6194225721784776,
+      "grad_norm": 0.3662350700773784,
+      "learning_rate": 5.189627262510655e-07,
+      "loss": 0.8695,
+      "step": 309
+    },
+    {
+      "epoch": 1.6246719160104988,
+      "grad_norm": 0.36967015418887056,
+      "learning_rate": 5.049569317994013e-07,
+      "loss": 0.8448,
+      "step": 310
+    },
+    {
+      "epoch": 1.6299212598425197,
+      "grad_norm": 0.3579476966594129,
+      "learning_rate": 4.911214673283157e-07,
+      "loss": 0.8375,
+      "step": 311
+    },
+    {
+      "epoch": 1.6351706036745406,
+      "grad_norm": 0.3611024965491316,
+      "learning_rate": 4.774575140626317e-07,
+      "loss": 0.8519,
+      "step": 312
+    },
+    {
+      "epoch": 1.6404199475065617,
+      "grad_norm": 0.3485470659035558,
+      "learning_rate": 4.639662385841293e-07,
+      "loss": 0.8217,
+      "step": 313
+    },
+    {
+      "epoch": 1.6456692913385826,
+      "grad_norm": 0.37287096492671606,
+      "learning_rate": 4.506487927319475e-07,
+      "loss": 0.8413,
+      "step": 314
+    },
+    {
+      "epoch": 1.6509186351706036,
+      "grad_norm": 0.38705573070810245,
+      "learning_rate": 4.3750631350424456e-07,
+      "loss": 0.8499,
+      "step": 315
+    },
+    {
+      "epoch": 1.6561679790026247,
+      "grad_norm": 0.3582574621798544,
+      "learning_rate": 4.2453992296112384e-07,
+      "loss": 0.8422,
+      "step": 316
+    },
+    {
+      "epoch": 1.6614173228346458,
+      "grad_norm": 0.36916485545424343,
+      "learning_rate": 4.117507281288366e-07,
+      "loss": 0.8312,
+      "step": 317
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.36367894508342596,
+      "learning_rate": 3.991398209052685e-07,
+      "loss": 0.8342,
+      "step": 318
+    },
+    {
+      "epoch": 1.6719160104986877,
+      "grad_norm": 0.39384159873675784,
+      "learning_rate": 3.8670827796671637e-07,
+      "loss": 0.8006,
+      "step": 319
+    },
+    {
+      "epoch": 1.6771653543307088,
+      "grad_norm": 0.39604366934016394,
+      "learning_rate": 3.7445716067596506e-07,
+      "loss": 0.8556,
+      "step": 320
+    },
+    {
+      "epoch": 1.6824146981627297,
+      "grad_norm": 0.3672901516398384,
+      "learning_rate": 3.623875149916725e-07,
+      "loss": 0.8621,
+      "step": 321
+    },
+    {
+      "epoch": 1.6876640419947506,
+      "grad_norm": 0.38389464950775926,
+      "learning_rate": 3.505003713790689e-07,
+      "loss": 0.8343,
+      "step": 322
+    },
+    {
+      "epoch": 1.6929133858267718,
+      "grad_norm": 0.36767539320644105,
+      "learning_rate": 3.387967447219803e-07,
+      "loss": 0.842,
+      "step": 323
+    },
+    {
+      "epoch": 1.6981627296587927,
+      "grad_norm": 0.3907752041858798,
+      "learning_rate": 3.2727763423617915e-07,
+      "loss": 0.8767,
+      "step": 324
+    },
+    {
+      "epoch": 1.7034120734908136,
+      "grad_norm": 0.37713933030901214,
+      "learning_rate": 3.1594402338407633e-07,
+      "loss": 0.8326,
+      "step": 325
+    },
+    {
+      "epoch": 1.7086614173228347,
+      "grad_norm": 0.38192173068300844,
+      "learning_rate": 3.047968797907552e-07,
+      "loss": 0.8168,
+      "step": 326
+    },
+    {
+      "epoch": 1.7139107611548556,
+      "grad_norm": 0.35632236660696437,
+      "learning_rate": 2.9383715516136083e-07,
+      "loss": 0.8344,
+      "step": 327
+    },
+    {
+      "epoch": 1.7191601049868765,
+      "grad_norm": 0.39281712398945773,
+      "learning_rate": 2.8306578519984526e-07,
+      "loss": 0.8057,
+      "step": 328
+    },
+    {
+      "epoch": 1.7244094488188977,
+      "grad_norm": 0.3713265865015316,
+      "learning_rate": 2.7248368952908055e-07,
+      "loss": 0.8128,
+      "step": 329
+    },
+    {
+      "epoch": 1.7296587926509186,
+      "grad_norm": 0.3704631773575278,
+      "learning_rate": 2.620917716123444e-07,
+      "loss": 0.8593,
+      "step": 330
+    },
+    {
+      "epoch": 1.7349081364829395,
+      "grad_norm": 0.3833563384401533,
+      "learning_rate": 2.5189091867618615e-07,
+      "loss": 0.8353,
+      "step": 331
+    },
+    {
+      "epoch": 1.7401574803149606,
+      "grad_norm": 0.39756909736931284,
+      "learning_rate": 2.418820016346779e-07,
+      "loss": 0.8596,
+      "step": 332
+    },
+    {
+      "epoch": 1.7454068241469818,
+      "grad_norm": 0.36212623044011794,
+      "learning_rate": 2.3206587501505866e-07,
+      "loss": 0.8311,
+      "step": 333
+    },
+    {
+      "epoch": 1.7506561679790025,
+      "grad_norm": 0.36098218836000906,
+      "learning_rate": 2.224433768847789e-07,
+      "loss": 0.8091,
+      "step": 334
+    },
+    {
+      "epoch": 1.7559055118110236,
+      "grad_norm": 0.3713293844163563,
+      "learning_rate": 2.1301532877994747e-07,
+      "loss": 0.8147,
+      "step": 335
+    },
+    {
+      "epoch": 1.7611548556430447,
+      "grad_norm": 0.34562968172819897,
+      "learning_rate": 2.0378253563519247e-07,
+      "loss": 0.8284,
+      "step": 336
+    },
+    {
+      "epoch": 1.7664041994750657,
+      "grad_norm": 0.39461152581923564,
+      "learning_rate": 1.9474578571493874e-07,
+      "loss": 0.8632,
+      "step": 337
+    },
+    {
+      "epoch": 1.7716535433070866,
+      "grad_norm": 0.3689417081438894,
+      "learning_rate": 1.859058505461095e-07,
+      "loss": 0.8259,
+      "step": 338
+    },
+    {
+      "epoch": 1.7769028871391077,
+      "grad_norm": 0.3926353728632124,
+      "learning_rate": 1.7726348485225337e-07,
+      "loss": 0.8364,
+      "step": 339
+    },
+    {
+      "epoch": 1.7821522309711286,
+      "grad_norm": 0.354022004999706,
+      "learning_rate": 1.6881942648911077e-07,
+      "loss": 0.8773,
+      "step": 340
+    },
+    {
+      "epoch": 1.7874015748031495,
+      "grad_norm": 0.3876142010040814,
+      "learning_rate": 1.6057439638161891e-07,
+      "loss": 0.82,
+      "step": 341
+    },
+    {
+      "epoch": 1.7926509186351707,
+      "grad_norm": 0.35474834193863947,
+      "learning_rate": 1.5252909846235898e-07,
+      "loss": 0.8193,
+      "step": 342
+    },
+    {
+      "epoch": 1.7979002624671916,
+      "grad_norm": 0.35916164129360717,
+      "learning_rate": 1.4468421961145924e-07,
+      "loss": 0.8398,
+      "step": 343
+    },
+    {
+      "epoch": 1.8031496062992125,
+      "grad_norm": 0.3592392678728242,
+      "learning_rate": 1.3704042959795132e-07,
+      "loss": 0.8384,
+      "step": 344
+    },
+    {
+      "epoch": 1.8083989501312336,
+      "grad_norm": 0.3717050715996407,
+      "learning_rate": 1.2959838102258537e-07,
+      "loss": 0.827,
+      "step": 345
+    },
+    {
+      "epoch": 1.8136482939632546,
+      "grad_norm": 0.3714676540830701,
+      "learning_rate": 1.223587092621162e-07,
+      "loss": 0.811,
+      "step": 346
+    },
+    {
+      "epoch": 1.8188976377952755,
+      "grad_norm": 0.36420858424939995,
+      "learning_rate": 1.1532203241505474e-07,
+      "loss": 0.8769,
+      "step": 347
+    },
+    {
+      "epoch": 1.8241469816272966,
+      "grad_norm": 0.3993728488042469,
+      "learning_rate": 1.0848895124889819e-07,
+      "loss": 0.8437,
+      "step": 348
+    },
+    {
+      "epoch": 1.8293963254593177,
+      "grad_norm": 0.36458761423411645,
+      "learning_rate": 1.0186004914883779e-07,
+      "loss": 0.8378,
+      "step": 349
+    },
+    {
+      "epoch": 1.8346456692913384,
+      "grad_norm": 0.3874822926899934,
+      "learning_rate": 9.54358920679524e-08,
+      "loss": 0.8284,
+      "step": 350
+    },
+    {
+      "epoch": 1.8398950131233596,
+      "grad_norm": 0.38256154200443654,
+      "learning_rate": 8.921702847888791e-08,
+      "loss": 0.8602,
+      "step": 351
+    },
+    {
+      "epoch": 1.8451443569553807,
+      "grad_norm": 0.36179666269244404,
+      "learning_rate": 8.320398932703145e-08,
+      "loss": 0.8274,
+      "step": 352
+    },
+    {
+      "epoch": 1.8503937007874016,
+      "grad_norm": 0.365807165050502,
+      "learning_rate": 7.739728798518115e-08,
+      "loss": 0.8709,
+      "step": 353
+    },
+    {
+      "epoch": 1.8556430446194225,
+      "grad_norm": 0.35458800620057285,
+      "learning_rate": 7.17974202097152e-08,
+      "loss": 0.8111,
+      "step": 354
+    },
+    {
+      "epoch": 1.8608923884514437,
+      "grad_norm": 0.37957333795063364,
+      "learning_rate": 6.640486409826785e-08,
+      "loss": 0.7994,
+      "step": 355
+    },
+    {
+      "epoch": 1.8661417322834646,
+      "grad_norm": 0.3527171222570775,
+      "learning_rate": 6.12200800489085e-08,
+      "loss": 0.853,
+      "step": 356
+    },
+    {
+      "epoch": 1.8713910761154855,
+      "grad_norm": 0.39242714191446193,
+      "learning_rate": 5.624351072083561e-08,
+      "loss": 0.8203,
+      "step": 357
+    },
+    {
+      "epoch": 1.8766404199475066,
+      "grad_norm": 0.35707175993984824,
+      "learning_rate": 5.1475580996585285e-08,
+      "loss": 0.7958,
+      "step": 358
+    },
+    {
+      "epoch": 1.8818897637795275,
+      "grad_norm": 0.3335686552470708,
+      "learning_rate": 4.691669794575388e-08,
+      "loss": 0.8433,
+      "step": 359
+    },
+    {
+      "epoch": 1.8871391076115485,
+      "grad_norm": 0.38462069858291403,
+      "learning_rate": 4.256725079024554e-08,
+      "loss": 0.7975,
+      "step": 360
+    },
+    {
+      "epoch": 1.8923884514435696,
+      "grad_norm": 0.34131290146851945,
+      "learning_rate": 3.8427610871041024e-08,
+      "loss": 0.8223,
+      "step": 361
+    },
+    {
+      "epoch": 1.8976377952755905,
+      "grad_norm": 0.35781280354041417,
+      "learning_rate": 3.449813161649357e-08,
+      "loss": 0.9063,
+      "step": 362
+    },
+    {
+      "epoch": 1.9028871391076114,
+      "grad_norm": 0.33992014890280325,
+      "learning_rate": 3.077914851215585e-08,
+      "loss": 0.8081,
+      "step": 363
+    },
+    {
+      "epoch": 1.9081364829396326,
+      "grad_norm": 0.3582145429279079,
+      "learning_rate": 2.7270979072135106e-08,
+      "loss": 0.8487,
+      "step": 364
+    },
+    {
+      "epoch": 1.9133858267716537,
+      "grad_norm": 0.36355378789809917,
+      "learning_rate": 2.3973922811987295e-08,
+      "loss": 0.8128,
+      "step": 365
+    },
+    {
+      "epoch": 1.9186351706036744,
+      "grad_norm": 0.36811282667120626,
+      "learning_rate": 2.0888261223143136e-08,
+      "loss": 0.8442,
+      "step": 366
+    },
+    {
+      "epoch": 1.9238845144356955,
+      "grad_norm": 0.36592765014539297,
+      "learning_rate": 1.8014257748877606e-08,
+      "loss": 0.8385,
+      "step": 367
+    },
+    {
+      "epoch": 1.9291338582677167,
+      "grad_norm": 0.3695353834129384,
+      "learning_rate": 1.5352157761815978e-08,
+      "loss": 0.809,
+      "step": 368
+    },
+    {
+      "epoch": 1.9343832020997376,
+      "grad_norm": 0.3907850295180158,
+      "learning_rate": 1.2902188542986139e-08,
+      "loss": 0.8295,
+      "step": 369
+    },
+    {
+      "epoch": 1.9396325459317585,
+      "grad_norm": 0.34933462418116523,
+      "learning_rate": 1.0664559262413831e-08,
+      "loss": 0.8238,
+      "step": 370
+    },
+    {
+      "epoch": 1.9448818897637796,
+      "grad_norm": 0.3537875486215043,
+      "learning_rate": 8.639460961263612e-09,
+      "loss": 0.8377,
+      "step": 371
+    },
+    {
+      "epoch": 1.9501312335958005,
+      "grad_norm": 0.35563927647325777,
+      "learning_rate": 6.827066535529947e-09,
+      "loss": 0.7943,
+      "step": 372
+    },
+    {
+      "epoch": 1.9553805774278215,
+      "grad_norm": 0.3474400330562922,
+      "learning_rate": 5.2275307212742986e-09,
+      "loss": 0.8511,
+      "step": 373
+    },
+    {
+      "epoch": 1.9606299212598426,
+      "grad_norm": 0.3593397542649248,
+      "learning_rate": 3.840990081415141e-09,
+      "loss": 0.8186,
+      "step": 374
+    },
+    {
+      "epoch": 1.9658792650918635,
+      "grad_norm": 0.3612580709923566,
+      "learning_rate": 2.6675629940689508e-09,
+      "loss": 0.848,
+      "step": 375
+    },
+    {
+      "epoch": 1.9711286089238844,
+      "grad_norm": 0.3614238458155997,
+      "learning_rate": 1.707349642442735e-09,
+      "loss": 0.7933,
+      "step": 376
+    },
+    {
+      "epoch": 1.9763779527559056,
+      "grad_norm": 0.39289720311660936,
+      "learning_rate": 9.604320062814309e-10,
+      "loss": 0.8323,
+      "step": 377
+    },
+    {
+      "epoch": 1.9816272965879265,
+      "grad_norm": 0.36167265662489084,
+      "learning_rate": 4.268738548682261e-10,
+      "loss": 0.8512,
+      "step": 378
+    },
+    {
+      "epoch": 1.9868766404199474,
+      "grad_norm": 0.3690750146631276,
+      "learning_rate": 1.0672074158030176e-10,
+      "loss": 0.832,
+      "step": 379
+    },
+    {
+      "epoch": 1.9921259842519685,
+      "grad_norm": 0.35408134628230764,
+      "learning_rate": 0.0,
+      "loss": 0.8172,
+      "step": 380
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 380,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 95,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.333503833071944e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4bd5857f3af9b6fbcb1b4a0bbae3be221ab574653fcbbd77e41a92872dabd19
+size 8440

zero_to_fp32.py

@@ -0,0 +1,760 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import gc
+import json
+import numpy as np
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device, weights_only=False)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in tqdm(files, desc='Loading checkpoint shards'):
+        state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+class GatheredTensor:
+    """
+    A pseudo tensor that collects partitioned weights.
+    It is more memory efficient when there are multiple groups.
+    """
+
+    def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape):
+        self.flat_groups = flat_groups
+        self.flat_groups_offset = flat_groups_offset
+        self.offset = offset
+        self.partitioned_numel = partitioned_numel
+        self.shape = shape
+        self.dtype = self.flat_groups[0][0].dtype
+
+    def contiguous(self):
+        """
+        Merge partitioned weights from flat_groups into a single tensor.
+        """
+        end_idx = self.offset + self.partitioned_numel
+        world_size = len(self.flat_groups)
+        pad_flat_param_chunks = []
+
+        for rank_i in range(world_size):
+            # for each rank, we need to collect weights from related group/groups
+            flat_groups_at_rank_i = self.flat_groups[rank_i]
+            start_group_id = None
+            end_group_id = None
+            for group_id in range(len(self.flat_groups_offset)):
+                if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]:
+                    start_group_id = group_id
+                if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]:
+                    end_group_id = group_id
+                    break
+            # collect weights from related group/groups
+            for group_id in range(start_group_id, end_group_id + 1):
+                flat_tensor = flat_groups_at_rank_i[group_id]
+                start_offset = self.offset - self.flat_groups_offset[group_id]
+                end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id]
+                pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset])
+
+        # collect weights from all ranks
+        pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0)
+        param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous()
+        return param
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * world_size
+
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]]))
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # memory efficient tensor
+        tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape)
+        state_dict[name] = tensor
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def to_torch_tensor(state_dict, return_empty_tensor=False):
+    """
+    Convert state_dict of GatheredTensor to torch tensor
+    """
+    torch_state_dict = {}
+    converted_tensors = {}
+    for name, tensor in state_dict.items():
+        tensor_id = id(tensor)
+        if tensor_id in converted_tensors:  # shared tensors
+            shared_tensor = torch_state_dict[converted_tensors[tensor_id]]
+            torch_state_dict[name] = shared_tensor
+        else:
+            converted_tensors[tensor_id] = name
+            if return_empty_tensor:
+                torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype)
+            else:
+                torch_state_dict[name] = tensor.contiguous()
+    return torch_state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                             tag=None,
+                                             exclude_frozen_parameters=False,
+                                             lazy_mode=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+        - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient.
+          Convert the pesduo tensor to torch tensor by ``.contiguous()``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    Note: the above usage may not work if your application doesn't have sufficient free CPU memory.
+    You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint. Or you can load state_dict in lazy mode ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu
+        for name, lazy_tensor in state_dict.item():
+            tensor = lazy_tensor.contiguous()  # to cpu
+            print(name, tensor)
+            # del tensor to release memory if it no longer in use
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+    if lazy_mode:
+        return state_dict
+    else:
+        return to_torch_tensor(state_dict)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                                          tag,
+                                                          exclude_frozen_parameters,
+                                                          lazy_mode=True)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        # an memory-efficient approach for sharding
+        empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True)
+        state_dict_split = split_torch_state_dict_into_shards(empty_state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model by shard
+    os.makedirs(output_dir, exist_ok=True)
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors}
+        shard_state_dict = to_torch_tensor(shard_state_dict)
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard_state_dict, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard_state_dict, output_path)
+        # release the memory of current shard
+        for tensor_name in list(shard_state_dict.keys()):
+            del state_dict[tensor_name]
+            del shard_state_dict[tensor_name]
+        del shard_state_dict
+        gc.collect()
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)