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": "./qwq",
+  "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:c4486be345bf9812138d2e70a5438f3d77a3277f50e292a592db44530eb5ed09
+size 4891730992

model-00002-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c0b7094520949e5db84534b10d4bfc2c2f50d6e8dd40841b0c13ff314225a5c
+size 4876059352

model-00003-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5c6188f06710445c82cf9fcb662dc5a4b107ddc1fa27f446f7d5f89417dce64
+size 4876059384

model-00004-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:781650d4985de841b70e4a0a344777f9a0beb18b22610aa6860ef56f748e9ae9
+size 4876059416

model-00005-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:219b1301d713217fe00d1d710d390a8fc0187a262c89dff248f8398433b2f1aa
+size 4876059416

model-00006-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce082af15ff53ea31f34ee410143b65e819e32580436ce8267bce60542b3fa84
+size 4876059416

model-00007-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ad6e09861e664b1f2fa0c39af92f1842bf2435944b1d8547585d77e5ad27d20
+size 4876059416

model-00008-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0480652b2bb01a90484b4c383c670ec7e081019a4fdace451e34a3497e96ad7
+size 4876059416

model-00009-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:529b82f39a5064cb4598bc71a7e00e3f1ac9911877034928c1f5a42c7a611814
+size 4876059416

model-00010-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da1e7c99bf03fbcb701a422bd51a6c5b6ec99e3f71e3922b415c9afdcc61f412
+size 4876059416

model-00011-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c03b8133dd219e4bbc5052ad02f75245d56d0a07d8cc3e70f590d81878f0e35
+size 4876059416

model-00012-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a8235fe12460e0611e8836f39c345deb90025a87b544c4619a2275aa531f8fe
+size 4876059416

model-00013-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16563b864450652c25cdb771aeaa0b212b7960b88b5bba64eb91681aa06cbdc1
+size 4876059416

model-00014-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4556c6da97da959950d67b8e281f405d20ab01ad7c585ace0013b39b89e45ce2
+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:d69387a9f094d5086cb662a4d68368958913df5d18c9a021df6806b5951e8bf0
+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": 1.1348930782232016,
+      "learning_rate": 1.5000000000000002e-07,
+      "loss": 1.1087,
+      "step": 1
+    },
+    {
+      "epoch": 0.010498687664041995,
+      "grad_norm": 1.123696373079589,
+      "learning_rate": 3.0000000000000004e-07,
+      "loss": 1.1356,
+      "step": 2
+    },
+    {
+      "epoch": 0.015748031496062992,
+      "grad_norm": 1.0989081863562118,
+      "learning_rate": 4.5e-07,
+      "loss": 1.1158,
+      "step": 3
+    },
+    {
+      "epoch": 0.02099737532808399,
+      "grad_norm": 1.0628548113414964,
+      "learning_rate": 6.000000000000001e-07,
+      "loss": 1.0986,
+      "step": 4
+    },
+    {
+      "epoch": 0.026246719160104987,
+      "grad_norm": 1.0629069543612368,
+      "learning_rate": 7.5e-07,
+      "loss": 1.0727,
+      "step": 5
+    },
+    {
+      "epoch": 0.031496062992125984,
+      "grad_norm": 1.1219311917213644,
+      "learning_rate": 9e-07,
+      "loss": 1.1513,
+      "step": 6
+    },
+    {
+      "epoch": 0.03674540682414698,
+      "grad_norm": 1.068318638334139,
+      "learning_rate": 1.05e-06,
+      "loss": 1.0978,
+      "step": 7
+    },
+    {
+      "epoch": 0.04199475065616798,
+      "grad_norm": 1.0335025624008565,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 1.0932,
+      "step": 8
+    },
+    {
+      "epoch": 0.047244094488188976,
+      "grad_norm": 0.9514112971268772,
+      "learning_rate": 1.35e-06,
+      "loss": 1.1046,
+      "step": 9
+    },
+    {
+      "epoch": 0.05249343832020997,
+      "grad_norm": 0.8944230714776324,
+      "learning_rate": 1.5e-06,
+      "loss": 1.0638,
+      "step": 10
+    },
+    {
+      "epoch": 0.05774278215223097,
+      "grad_norm": 0.8720343077794245,
+      "learning_rate": 1.65e-06,
+      "loss": 1.1132,
+      "step": 11
+    },
+    {
+      "epoch": 0.06299212598425197,
+      "grad_norm": 0.7519518665820406,
+      "learning_rate": 1.8e-06,
+      "loss": 1.0788,
+      "step": 12
+    },
+    {
+      "epoch": 0.06824146981627296,
+      "grad_norm": 0.7768466543241798,
+      "learning_rate": 1.95e-06,
+      "loss": 1.0795,
+      "step": 13
+    },
+    {
+      "epoch": 0.07349081364829396,
+      "grad_norm": 0.7109922479048013,
+      "learning_rate": 2.1e-06,
+      "loss": 1.1012,
+      "step": 14
+    },
+    {
+      "epoch": 0.07874015748031496,
+      "grad_norm": 0.6312078880187205,
+      "learning_rate": 2.25e-06,
+      "loss": 1.0851,
+      "step": 15
+    },
+    {
+      "epoch": 0.08398950131233596,
+      "grad_norm": 0.5514473048370377,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 1.1041,
+      "step": 16
+    },
+    {
+      "epoch": 0.08923884514435695,
+      "grad_norm": 0.6271281070432462,
+      "learning_rate": 2.55e-06,
+      "loss": 1.0855,
+      "step": 17
+    },
+    {
+      "epoch": 0.09448818897637795,
+      "grad_norm": 0.7059888078645049,
+      "learning_rate": 2.7e-06,
+      "loss": 1.0473,
+      "step": 18
+    },
+    {
+      "epoch": 0.09973753280839895,
+      "grad_norm": 0.7226157330393405,
+      "learning_rate": 2.85e-06,
+      "loss": 1.0665,
+      "step": 19
+    },
+    {
+      "epoch": 0.10498687664041995,
+      "grad_norm": 0.7244742832208652,
+      "learning_rate": 3e-06,
+      "loss": 1.0604,
+      "step": 20
+    },
+    {
+      "epoch": 0.11023622047244094,
+      "grad_norm": 0.7088251146482789,
+      "learning_rate": 3.1500000000000003e-06,
+      "loss": 1.0516,
+      "step": 21
+    },
+    {
+      "epoch": 0.11548556430446194,
+      "grad_norm": 0.5987242362229293,
+      "learning_rate": 3.3e-06,
+      "loss": 1.084,
+      "step": 22
+    },
+    {
+      "epoch": 0.12073490813648294,
+      "grad_norm": 0.5730637810768702,
+      "learning_rate": 3.45e-06,
+      "loss": 1.0621,
+      "step": 23
+    },
+    {
+      "epoch": 0.12598425196850394,
+      "grad_norm": 0.5894968443138215,
+      "learning_rate": 3.6e-06,
+      "loss": 1.0797,
+      "step": 24
+    },
+    {
+      "epoch": 0.13123359580052493,
+      "grad_norm": 0.5798124303184627,
+      "learning_rate": 3.75e-06,
+      "loss": 1.0035,
+      "step": 25
+    },
+    {
+      "epoch": 0.13648293963254593,
+      "grad_norm": 0.643205751513686,
+      "learning_rate": 3.9e-06,
+      "loss": 1.0455,
+      "step": 26
+    },
+    {
+      "epoch": 0.14173228346456693,
+      "grad_norm": 0.5621970774702022,
+      "learning_rate": 4.05e-06,
+      "loss": 1.0576,
+      "step": 27
+    },
+    {
+      "epoch": 0.14698162729658792,
+      "grad_norm": 0.5506084571895594,
+      "learning_rate": 4.2e-06,
+      "loss": 1.0298,
+      "step": 28
+    },
+    {
+      "epoch": 0.15223097112860892,
+      "grad_norm": 0.48741149421912777,
+      "learning_rate": 4.35e-06,
+      "loss": 1.0018,
+      "step": 29
+    },
+    {
+      "epoch": 0.15748031496062992,
+      "grad_norm": 0.46403007703544275,
+      "learning_rate": 4.5e-06,
+      "loss": 0.9872,
+      "step": 30
+    },
+    {
+      "epoch": 0.16272965879265092,
+      "grad_norm": 0.4754381818573106,
+      "learning_rate": 4.65e-06,
+      "loss": 1.0271,
+      "step": 31
+    },
+    {
+      "epoch": 0.1679790026246719,
+      "grad_norm": 0.9362850890979981,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 1.0437,
+      "step": 32
+    },
+    {
+      "epoch": 0.1732283464566929,
+      "grad_norm": 0.47391181595772164,
+      "learning_rate": 4.95e-06,
+      "loss": 1.0437,
+      "step": 33
+    },
+    {
+      "epoch": 0.1784776902887139,
+      "grad_norm": 0.5276920454851337,
+      "learning_rate": 5.1e-06,
+      "loss": 1.0557,
+      "step": 34
+    },
+    {
+      "epoch": 0.1837270341207349,
+      "grad_norm": 0.4616075133913133,
+      "learning_rate": 5.2500000000000006e-06,
+      "loss": 1.0465,
+      "step": 35
+    },
+    {
+      "epoch": 0.1889763779527559,
+      "grad_norm": 0.4555174555636226,
+      "learning_rate": 5.4e-06,
+      "loss": 1.0588,
+      "step": 36
+    },
+    {
+      "epoch": 0.1942257217847769,
+      "grad_norm": 0.5071864534648831,
+      "learning_rate": 5.55e-06,
+      "loss": 1.044,
+      "step": 37
+    },
+    {
+      "epoch": 0.1994750656167979,
+      "grad_norm": 0.4851367263882934,
+      "learning_rate": 5.7e-06,
+      "loss": 1.0464,
+      "step": 38
+    },
+    {
+      "epoch": 0.2047244094488189,
+      "grad_norm": 0.44188022228811896,
+      "learning_rate": 5.85e-06,
+      "loss": 1.0182,
+      "step": 39
+    },
+    {
+      "epoch": 0.2099737532808399,
+      "grad_norm": 0.43420740120454643,
+      "learning_rate": 6e-06,
+      "loss": 1.0188,
+      "step": 40
+    },
+    {
+      "epoch": 0.2152230971128609,
+      "grad_norm": 0.4291543441241407,
+      "learning_rate": 5.9998719351101036e-06,
+      "loss": 1.0245,
+      "step": 41
+    },
+    {
+      "epoch": 0.2204724409448819,
+      "grad_norm": 0.43326370236005163,
+      "learning_rate": 5.999487751374158e-06,
+      "loss": 1.0238,
+      "step": 42
+    },
+    {
+      "epoch": 0.22572178477690288,
+      "grad_norm": 0.427571644972227,
+      "learning_rate": 5.998847481592462e-06,
+      "loss": 1.0311,
+      "step": 43
+    },
+    {
+      "epoch": 0.23097112860892388,
+      "grad_norm": 0.4215063088273006,
+      "learning_rate": 5.997951180429069e-06,
+      "loss": 0.9925,
+      "step": 44
+    },
+    {
+      "epoch": 0.23622047244094488,
+      "grad_norm": 0.4206536914503675,
+      "learning_rate": 5.996798924407118e-06,
+      "loss": 1.003,
+      "step": 45
+    },
+    {
+      "epoch": 0.24146981627296588,
+      "grad_norm": 0.40910969064965136,
+      "learning_rate": 5.995390811902302e-06,
+      "loss": 0.9949,
+      "step": 46
+    },
+    {
+      "epoch": 0.24671916010498687,
+      "grad_norm": 0.4165775049327623,
+      "learning_rate": 5.993726963134471e-06,
+      "loss": 0.9734,
+      "step": 47
+    },
+    {
+      "epoch": 0.25196850393700787,
+      "grad_norm": 0.3832235501001726,
+      "learning_rate": 5.9918075201573645e-06,
+      "loss": 0.9485,
+      "step": 48
+    },
+    {
+      "epoch": 0.2572178477690289,
+      "grad_norm": 0.37002495168808525,
+      "learning_rate": 5.9896326468464835e-06,
+      "loss": 0.9358,
+      "step": 49
+    },
+    {
+      "epoch": 0.26246719160104987,
+      "grad_norm": 0.44836853406053057,
+      "learning_rate": 5.987202528885104e-06,
+      "loss": 0.9982,
+      "step": 50
+    },
+    {
+      "epoch": 0.2677165354330709,
+      "grad_norm": 0.4080608606117312,
+      "learning_rate": 5.984517373748417e-06,
+      "loss": 1.0129,
+      "step": 51
+    },
+    {
+      "epoch": 0.27296587926509186,
+      "grad_norm": 0.4001550595702573,
+      "learning_rate": 5.981577410685822e-06,
+      "loss": 0.9788,
+      "step": 52
+    },
+    {
+      "epoch": 0.2782152230971129,
+      "grad_norm": 0.41021488877460305,
+      "learning_rate": 5.978382890701347e-06,
+      "loss": 1.0262,
+      "step": 53
+    },
+    {
+      "epoch": 0.28346456692913385,
+      "grad_norm": 0.39997016380492506,
+      "learning_rate": 5.9749340865322284e-06,
+      "loss": 1.0275,
+      "step": 54
+    },
+    {
+      "epoch": 0.2887139107611549,
+      "grad_norm": 0.3839823787027912,
+      "learning_rate": 5.971231292625615e-06,
+      "loss": 0.9374,
+      "step": 55
+    },
+    {
+      "epoch": 0.29396325459317585,
+      "grad_norm": 0.4125068495663659,
+      "learning_rate": 5.967274825113438e-06,
+      "loss": 0.9954,
+      "step": 56
+    },
+    {
+      "epoch": 0.2992125984251969,
+      "grad_norm": 0.3908377197765856,
+      "learning_rate": 5.963065021785414e-06,
+      "loss": 0.9671,
+      "step": 57
+    },
+    {
+      "epoch": 0.30446194225721784,
+      "grad_norm": 0.3850488592862481,
+      "learning_rate": 5.958602242060207e-06,
+      "loss": 0.9657,
+      "step": 58
+    },
+    {
+      "epoch": 0.30971128608923887,
+      "grad_norm": 0.3877990366088493,
+      "learning_rate": 5.95388686695475e-06,
+      "loss": 0.9678,
+      "step": 59
+    },
+    {
+      "epoch": 0.31496062992125984,
+      "grad_norm": 0.40470471194287355,
+      "learning_rate": 5.948919299051706e-06,
+      "loss": 1.0149,
+      "step": 60
+    },
+    {
+      "epoch": 0.32020997375328086,
+      "grad_norm": 0.42889495063392963,
+      "learning_rate": 5.943699962465096e-06,
+      "loss": 1.033,
+      "step": 61
+    },
+    {
+      "epoch": 0.32545931758530183,
+      "grad_norm": 0.39164358737100274,
+      "learning_rate": 5.9382293028040985e-06,
+      "loss": 0.9761,
+      "step": 62
+    },
+    {
+      "epoch": 0.33070866141732286,
+      "grad_norm": 0.3869342590567232,
+      "learning_rate": 5.9325077871349975e-06,
+      "loss": 0.9982,
+      "step": 63
+    },
+    {
+      "epoch": 0.3359580052493438,
+      "grad_norm": 0.39264627926569035,
+      "learning_rate": 5.9265359039413105e-06,
+      "loss": 0.9667,
+      "step": 64
+    },
+    {
+      "epoch": 0.34120734908136485,
+      "grad_norm": 0.3887717698297268,
+      "learning_rate": 5.920314163082079e-06,
+      "loss": 0.9806,
+      "step": 65
+    },
+    {
+      "epoch": 0.3464566929133858,
+      "grad_norm": 0.40896336915084297,
+      "learning_rate": 5.913843095748342e-06,
+      "loss": 1.0135,
+      "step": 66
+    },
+    {
+      "epoch": 0.35170603674540685,
+      "grad_norm": 0.3610209560875707,
+      "learning_rate": 5.907123254417783e-06,
+      "loss": 0.956,
+      "step": 67
+    },
+    {
+      "epoch": 0.3569553805774278,
+      "grad_norm": 0.38154744815823505,
+      "learning_rate": 5.9001552128075625e-06,
+      "loss": 1.0045,
+      "step": 68
+    },
+    {
+      "epoch": 0.36220472440944884,
+      "grad_norm": 0.4094826396119445,
+      "learning_rate": 5.892939565825335e-06,
+      "loss": 1.0069,
+      "step": 69
+    },
+    {
+      "epoch": 0.3674540682414698,
+      "grad_norm": 0.39129138622932325,
+      "learning_rate": 5.885476929518457e-06,
+      "loss": 0.9525,
+      "step": 70
+    },
+    {
+      "epoch": 0.37270341207349084,
+      "grad_norm": 0.3712890701175899,
+      "learning_rate": 5.8777679410213956e-06,
+      "loss": 0.9792,
+      "step": 71
+    },
+    {
+      "epoch": 0.3779527559055118,
+      "grad_norm": 0.4086264062600148,
+      "learning_rate": 5.869813258501323e-06,
+      "loss": 0.9926,
+      "step": 72
+    },
+    {
+      "epoch": 0.38320209973753283,
+      "grad_norm": 0.368975878599487,
+      "learning_rate": 5.861613561101934e-06,
+      "loss": 0.9643,
+      "step": 73
+    },
+    {
+      "epoch": 0.3884514435695538,
+      "grad_norm": 0.36792811629461203,
+      "learning_rate": 5.853169548885461e-06,
+      "loss": 0.9867,
+      "step": 74
+    },
+    {
+      "epoch": 0.3937007874015748,
+      "grad_norm": 0.3566251893981936,
+      "learning_rate": 5.844481942772898e-06,
+      "loss": 1.0069,
+      "step": 75
+    },
+    {
+      "epoch": 0.3989501312335958,
+      "grad_norm": 0.4578529359685586,
+      "learning_rate": 5.835551484482459e-06,
+      "loss": 1.0173,
+      "step": 76
+    },
+    {
+      "epoch": 0.4041994750656168,
+      "grad_norm": 0.3935925285922137,
+      "learning_rate": 5.826378936466249e-06,
+      "loss": 0.9743,
+      "step": 77
+    },
+    {
+      "epoch": 0.4094488188976378,
+      "grad_norm": 0.4109939217838428,
+      "learning_rate": 5.81696508184517e-06,
+      "loss": 0.9866,
+      "step": 78
+    },
+    {
+      "epoch": 0.4146981627296588,
+      "grad_norm": 0.3839870332489822,
+      "learning_rate": 5.807310724342058e-06,
+      "loss": 0.9516,
+      "step": 79
+    },
+    {
+      "epoch": 0.4199475065616798,
+      "grad_norm": 0.3774576797883406,
+      "learning_rate": 5.797416688213067e-06,
+      "loss": 0.9895,
+      "step": 80
+    },
+    {
+      "epoch": 0.4251968503937008,
+      "grad_norm": 0.3817468964498129,
+      "learning_rate": 5.787283818177297e-06,
+      "loss": 0.9632,
+      "step": 81
+    },
+    {
+      "epoch": 0.4304461942257218,
+      "grad_norm": 0.60843002346461,
+      "learning_rate": 5.776912979344669e-06,
+      "loss": 1.0166,
+      "step": 82
+    },
+    {
+      "epoch": 0.4356955380577428,
+      "grad_norm": 0.3858713700245362,
+      "learning_rate": 5.766305057142073e-06,
+      "loss": 0.9976,
+      "step": 83
+    },
+    {
+      "epoch": 0.4409448818897638,
+      "grad_norm": 0.3724153436541016,
+      "learning_rate": 5.755460957237769e-06,
+      "loss": 0.9645,
+      "step": 84
+    },
+    {
+      "epoch": 0.4461942257217848,
+      "grad_norm": 0.38201105695018567,
+      "learning_rate": 5.744381605464064e-06,
+      "loss": 0.9899,
+      "step": 85
+    },
+    {
+      "epoch": 0.45144356955380577,
+      "grad_norm": 0.38383930861007165,
+      "learning_rate": 5.7330679477382655e-06,
+      "loss": 0.9919,
+      "step": 86
+    },
+    {
+      "epoch": 0.4566929133858268,
+      "grad_norm": 0.4078870418259581,
+      "learning_rate": 5.7215209499819296e-06,
+      "loss": 0.9797,
+      "step": 87
+    },
+    {
+      "epoch": 0.46194225721784776,
+      "grad_norm": 0.38463767466523974,
+      "learning_rate": 5.709741598038387e-06,
+      "loss": 0.9597,
+      "step": 88
+    },
+    {
+      "epoch": 0.4671916010498688,
+      "grad_norm": 0.36309855116472584,
+      "learning_rate": 5.697730897588577e-06,
+      "loss": 0.9737,
+      "step": 89
+    },
+    {
+      "epoch": 0.47244094488188976,
+      "grad_norm": 0.4106701446638758,
+      "learning_rate": 5.685489874065187e-06,
+      "loss": 0.9683,
+      "step": 90
+    },
+    {
+      "epoch": 0.4776902887139108,
+      "grad_norm": 0.37110409255145443,
+      "learning_rate": 5.673019572565103e-06,
+      "loss": 1.0418,
+      "step": 91
+    },
+    {
+      "epoch": 0.48293963254593175,
+      "grad_norm": 0.3558357783330656,
+      "learning_rate": 5.660321057760186e-06,
+      "loss": 1.0055,
+      "step": 92
+    },
+    {
+      "epoch": 0.4881889763779528,
+      "grad_norm": 0.40499489938404787,
+      "learning_rate": 5.6473954138063674e-06,
+      "loss": 1.0113,
+      "step": 93
+    },
+    {
+      "epoch": 0.49343832020997375,
+      "grad_norm": 0.39428526462199764,
+      "learning_rate": 5.634243744251094e-06,
+      "loss": 0.9875,
+      "step": 94
+    },
+    {
+      "epoch": 0.49868766404199477,
+      "grad_norm": 0.3711741011240413,
+      "learning_rate": 5.620867171939109e-06,
+      "loss": 0.9749,
+      "step": 95
+    },
+    {
+      "epoch": 0.5039370078740157,
+      "grad_norm": 0.3961340085644134,
+      "learning_rate": 5.607266838916585e-06,
+      "loss": 0.982,
+      "step": 96
+    },
+    {
+      "epoch": 0.5091863517060368,
+      "grad_norm": 0.3784646685814138,
+      "learning_rate": 5.593443906333624e-06,
+      "loss": 0.9957,
+      "step": 97
+    },
+    {
+      "epoch": 0.5144356955380578,
+      "grad_norm": 0.3750460397069026,
+      "learning_rate": 5.579399554345118e-06,
+      "loss": 0.9755,
+      "step": 98
+    },
+    {
+      "epoch": 0.5196850393700787,
+      "grad_norm": 0.3746718538274792,
+      "learning_rate": 5.565134982009994e-06,
+      "loss": 0.9736,
+      "step": 99
+    },
+    {
+      "epoch": 0.5249343832020997,
+      "grad_norm": 0.38418890409196027,
+      "learning_rate": 5.550651407188843e-06,
+      "loss": 0.9506,
+      "step": 100
+    },
+    {
+      "epoch": 0.5301837270341208,
+      "grad_norm": 0.422976375435725,
+      "learning_rate": 5.535950066439941e-06,
+      "loss": 1.0141,
+      "step": 101
+    },
+    {
+      "epoch": 0.5354330708661418,
+      "grad_norm": 0.38354451243133536,
+      "learning_rate": 5.521032214913679e-06,
+      "loss": 0.9618,
+      "step": 102
+    },
+    {
+      "epoch": 0.5406824146981627,
+      "grad_norm": 0.38257660011773076,
+      "learning_rate": 5.505899126245397e-06,
+      "loss": 0.939,
+      "step": 103
+    },
+    {
+      "epoch": 0.5459317585301837,
+      "grad_norm": 0.3768438915225408,
+      "learning_rate": 5.490552092446652e-06,
+      "loss": 0.9675,
+      "step": 104
+    },
+    {
+      "epoch": 0.5511811023622047,
+      "grad_norm": 0.3749655286727107,
+      "learning_rate": 5.474992423794907e-06,
+      "loss": 0.9592,
+      "step": 105
+    },
+    {
+      "epoch": 0.5564304461942258,
+      "grad_norm": 0.38461916993489687,
+      "learning_rate": 5.459221448721664e-06,
+      "loss": 0.9623,
+      "step": 106
+    },
+    {
+      "epoch": 0.5616797900262467,
+      "grad_norm": 0.35648642966931204,
+      "learning_rate": 5.443240513699045e-06,
+      "loss": 0.985,
+      "step": 107
+    },
+    {
+      "epoch": 0.5669291338582677,
+      "grad_norm": 0.4051560712719681,
+      "learning_rate": 5.427050983124842e-06,
+      "loss": 0.9407,
+      "step": 108
+    },
+    {
+      "epoch": 0.5721784776902887,
+      "grad_norm": 0.3769879713701903,
+      "learning_rate": 5.410654239206021e-06,
+      "loss": 0.968,
+      "step": 109
+    },
+    {
+      "epoch": 0.5774278215223098,
+      "grad_norm": 0.3746822083724367,
+      "learning_rate": 5.394051681840719e-06,
+      "loss": 0.9497,
+      "step": 110
+    },
+    {
+      "epoch": 0.5826771653543307,
+      "grad_norm": 0.3987231911136733,
+      "learning_rate": 5.3772447284987216e-06,
+      "loss": 0.961,
+      "step": 111
+    },
+    {
+      "epoch": 0.5879265091863517,
+      "grad_norm": 0.37848222525971176,
+      "learning_rate": 5.36023481410045e-06,
+      "loss": 0.9707,
+      "step": 112
+    },
+    {
+      "epoch": 0.5931758530183727,
+      "grad_norm": 0.3794904855253974,
+      "learning_rate": 5.343023390894446e-06,
+      "loss": 0.9714,
+      "step": 113
+    },
+    {
+      "epoch": 0.5984251968503937,
+      "grad_norm": 0.37452267525256994,
+      "learning_rate": 5.325611928333389e-06,
+      "loss": 0.9406,
+      "step": 114
+    },
+    {
+      "epoch": 0.6036745406824147,
+      "grad_norm": 0.39474437059829304,
+      "learning_rate": 5.308001912948637e-06,
+      "loss": 0.9626,
+      "step": 115
+    },
+    {
+      "epoch": 0.6089238845144357,
+      "grad_norm": 0.4023921986663554,
+      "learning_rate": 5.290194848223309e-06,
+      "loss": 0.9889,
+      "step": 116
+    },
+    {
+      "epoch": 0.6141732283464567,
+      "grad_norm": 0.39963771712171875,
+      "learning_rate": 5.272192254463929e-06,
+      "loss": 0.9639,
+      "step": 117
+    },
+    {
+      "epoch": 0.6194225721784777,
+      "grad_norm": 0.3893586064595733,
+      "learning_rate": 5.2539956686706205e-06,
+      "loss": 0.9469,
+      "step": 118
+    },
+    {
+      "epoch": 0.6246719160104987,
+      "grad_norm": 0.4651495625439333,
+      "learning_rate": 5.2356066444058875e-06,
+      "loss": 0.9658,
+      "step": 119
+    },
+    {
+      "epoch": 0.6299212598425197,
+      "grad_norm": 0.39599728107932586,
+      "learning_rate": 5.217026751661978e-06,
+      "loss": 1.0137,
+      "step": 120
+    },
+    {
+      "epoch": 0.6351706036745407,
+      "grad_norm": 0.406988761369817,
+      "learning_rate": 5.198257576726835e-06,
+      "loss": 0.9306,
+      "step": 121
+    },
+    {
+      "epoch": 0.6404199475065617,
+      "grad_norm": 0.3611939094322339,
+      "learning_rate": 5.179300722048673e-06,
+      "loss": 0.9462,
+      "step": 122
+    },
+    {
+      "epoch": 0.6456692913385826,
+      "grad_norm": 0.3809841775392484,
+      "learning_rate": 5.1601578060991645e-06,
+      "loss": 0.953,
+      "step": 123
+    },
+    {
+      "epoch": 0.6509186351706037,
+      "grad_norm": 0.46022843064705843,
+      "learning_rate": 5.1408304632352575e-06,
+      "loss": 0.9422,
+      "step": 124
+    },
+    {
+      "epoch": 0.6561679790026247,
+      "grad_norm": 0.3979704646560941,
+      "learning_rate": 5.1213203435596425e-06,
+      "loss": 0.9751,
+      "step": 125
+    },
+    {
+      "epoch": 0.6614173228346457,
+      "grad_norm": 0.39388496260457084,
+      "learning_rate": 5.101629112779873e-06,
+      "loss": 0.9722,
+      "step": 126
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.3899148438115094,
+      "learning_rate": 5.08175845206615e-06,
+      "loss": 0.9652,
+      "step": 127
+    },
+    {
+      "epoch": 0.6719160104986877,
+      "grad_norm": 0.37391882787694275,
+      "learning_rate": 5.061710057907788e-06,
+      "loss": 0.9621,
+      "step": 128
+    },
+    {
+      "epoch": 0.6771653543307087,
+      "grad_norm": 0.39500875865406576,
+      "learning_rate": 5.041485641968385e-06,
+      "loss": 0.9899,
+      "step": 129
+    },
+    {
+      "epoch": 0.6824146981627297,
+      "grad_norm": 0.37540362490802714,
+      "learning_rate": 5.021086930939672e-06,
+      "loss": 0.9472,
+      "step": 130
+    },
+    {
+      "epoch": 0.6876640419947506,
+      "grad_norm": 0.3940788728379769,
+      "learning_rate": 5.000515666394105e-06,
+      "loss": 0.9479,
+      "step": 131
+    },
+    {
+      "epoch": 0.6929133858267716,
+      "grad_norm": 0.3919125365655477,
+      "learning_rate": 4.979773604636169e-06,
+      "loss": 0.9624,
+      "step": 132
+    },
+    {
+      "epoch": 0.6981627296587927,
+      "grad_norm": 0.3804552314744538,
+      "learning_rate": 4.958862516552433e-06,
+      "loss": 0.9806,
+      "step": 133
+    },
+    {
+      "epoch": 0.7034120734908137,
+      "grad_norm": 0.3674434286105591,
+      "learning_rate": 4.937784187460362e-06,
+      "loss": 0.9511,
+      "step": 134
+    },
+    {
+      "epoch": 0.7086614173228346,
+      "grad_norm": 0.4109777494732396,
+      "learning_rate": 4.916540416955884e-06,
+      "loss": 0.9943,
+      "step": 135
+    },
+    {
+      "epoch": 0.7139107611548556,
+      "grad_norm": 0.40231567788837497,
+      "learning_rate": 4.895133018759753e-06,
+      "loss": 0.9798,
+      "step": 136
+    },
+    {
+      "epoch": 0.7191601049868767,
+      "grad_norm": 0.3721834479908975,
+      "learning_rate": 4.873563820562698e-06,
+      "loss": 0.9504,
+      "step": 137
+    },
+    {
+      "epoch": 0.7244094488188977,
+      "grad_norm": 0.36127526200518306,
+      "learning_rate": 4.851834663869379e-06,
+      "loss": 0.9517,
+      "step": 138
+    },
+    {
+      "epoch": 0.7296587926509186,
+      "grad_norm": 0.3513827139135777,
+      "learning_rate": 4.82994740384117e-06,
+      "loss": 0.9835,
+      "step": 139
+    },
+    {
+      "epoch": 0.7349081364829396,
+      "grad_norm": 0.36760728272750326,
+      "learning_rate": 4.80790390913777e-06,
+      "loss": 0.9503,
+      "step": 140
+    },
+    {
+      "epoch": 0.7401574803149606,
+      "grad_norm": 0.36275280721999276,
+      "learning_rate": 4.785706061757656e-06,
+      "loss": 0.9743,
+      "step": 141
+    },
+    {
+      "epoch": 0.7454068241469817,
+      "grad_norm": 0.3733380512329921,
+      "learning_rate": 4.763355756877419e-06,
+      "loss": 0.9384,
+      "step": 142
+    },
+    {
+      "epoch": 0.7506561679790026,
+      "grad_norm": 0.3801691027568987,
+      "learning_rate": 4.740854902689947e-06,
+      "loss": 0.9296,
+      "step": 143
+    },
+    {
+      "epoch": 0.7559055118110236,
+      "grad_norm": 0.39053906811778566,
+      "learning_rate": 4.718205420241516e-06,
+      "loss": 0.9488,
+      "step": 144
+    },
+    {
+      "epoch": 0.7611548556430446,
+      "grad_norm": 0.3923993707534958,
+      "learning_rate": 4.695409243267776e-06,
+      "loss": 0.9383,
+      "step": 145
+    },
+    {
+      "epoch": 0.7664041994750657,
+      "grad_norm": 0.364792552828712,
+      "learning_rate": 4.672468318028657e-06,
+      "loss": 0.9193,
+      "step": 146
+    },
+    {
+      "epoch": 0.7716535433070866,
+      "grad_norm": 0.35070825551906964,
+      "learning_rate": 4.649384603142202e-06,
+      "loss": 0.9164,
+      "step": 147
+    },
+    {
+      "epoch": 0.7769028871391076,
+      "grad_norm": 0.37099778180795795,
+      "learning_rate": 4.626160069417348e-06,
+      "loss": 0.9425,
+      "step": 148
+    },
+    {
+      "epoch": 0.7821522309711286,
+      "grad_norm": 0.36954118968922517,
+      "learning_rate": 4.602796699685665e-06,
+      "loss": 0.9265,
+      "step": 149
+    },
+    {
+      "epoch": 0.7874015748031497,
+      "grad_norm": 0.4076466706382121,
+      "learning_rate": 4.579296488632067e-06,
+      "loss": 1.0133,
+      "step": 150
+    },
+    {
+      "epoch": 0.7926509186351706,
+      "grad_norm": 0.4015334925568992,
+      "learning_rate": 4.5556614426245165e-06,
+      "loss": 0.9486,
+      "step": 151
+    },
+    {
+      "epoch": 0.7979002624671916,
+      "grad_norm": 0.39628644809730684,
+      "learning_rate": 4.5318935795427206e-06,
+      "loss": 0.9605,
+      "step": 152
+    },
+    {
+      "epoch": 0.8031496062992126,
+      "grad_norm": 0.36792154742540445,
+      "learning_rate": 4.507994928605862e-06,
+      "loss": 0.9287,
+      "step": 153
+    },
+    {
+      "epoch": 0.8083989501312336,
+      "grad_norm": 0.3887839296706913,
+      "learning_rate": 4.483967530199337e-06,
+      "loss": 0.951,
+      "step": 154
+    },
+    {
+      "epoch": 0.8136482939632546,
+      "grad_norm": 0.36716852968968616,
+      "learning_rate": 4.459813435700569e-06,
+      "loss": 0.9702,
+      "step": 155
+    },
+    {
+      "epoch": 0.8188976377952756,
+      "grad_norm": 0.3533521076976156,
+      "learning_rate": 4.4355347073038595e-06,
+      "loss": 0.9612,
+      "step": 156
+    },
+    {
+      "epoch": 0.8241469816272966,
+      "grad_norm": 0.3499649930079787,
+      "learning_rate": 4.411133417844328e-06,
+      "loss": 0.9599,
+      "step": 157
+    },
+    {
+      "epoch": 0.8293963254593176,
+      "grad_norm": 0.38582146832565867,
+      "learning_rate": 4.38661165062094e-06,
+      "loss": 0.9894,
+      "step": 158
+    },
+    {
+      "epoch": 0.8346456692913385,
+      "grad_norm": 0.39040836855795735,
+      "learning_rate": 4.36197149921864e-06,
+      "loss": 0.9747,
+      "step": 159
+    },
+    {
+      "epoch": 0.8398950131233596,
+      "grad_norm": 0.3798580758700489,
+      "learning_rate": 4.3372150673296155e-06,
+      "loss": 0.9654,
+      "step": 160
+    },
+    {
+      "epoch": 0.8451443569553806,
+      "grad_norm": 0.3764456540061034,
+      "learning_rate": 4.3123444685736795e-06,
+      "loss": 0.9823,
+      "step": 161
+    },
+    {
+      "epoch": 0.8503937007874016,
+      "grad_norm": 0.3771195417830333,
+      "learning_rate": 4.287361826317827e-06,
+      "loss": 0.9456,
+      "step": 162
+    },
+    {
+      "epoch": 0.8556430446194225,
+      "grad_norm": 0.37650137746409273,
+      "learning_rate": 4.262269273494946e-06,
+      "loss": 1.0022,
+      "step": 163
+    },
+    {
+      "epoch": 0.8608923884514436,
+      "grad_norm": 0.38148353077474145,
+      "learning_rate": 4.237068952421711e-06,
+      "loss": 0.964,
+      "step": 164
+    },
+    {
+      "epoch": 0.8661417322834646,
+      "grad_norm": 0.3982519128695332,
+      "learning_rate": 4.2117630146156845e-06,
+      "loss": 0.9673,
+      "step": 165
+    },
+    {
+      "epoch": 0.8713910761154856,
+      "grad_norm": 0.36000775624632003,
+      "learning_rate": 4.186353620611627e-06,
+      "loss": 0.9359,
+      "step": 166
+    },
+    {
+      "epoch": 0.8766404199475065,
+      "grad_norm": 0.36850454735662447,
+      "learning_rate": 4.160842939777036e-06,
+      "loss": 0.9422,
+      "step": 167
+    },
+    {
+      "epoch": 0.8818897637795275,
+      "grad_norm": 0.37804115639757085,
+      "learning_rate": 4.135233150126931e-06,
+      "loss": 0.9454,
+      "step": 168
+    },
+    {
+      "epoch": 0.8871391076115486,
+      "grad_norm": 0.3689383402086321,
+      "learning_rate": 4.109526438137908e-06,
+      "loss": 0.9455,
+      "step": 169
+    },
+    {
+      "epoch": 0.8923884514435696,
+      "grad_norm": 0.46527154775209717,
+      "learning_rate": 4.08372499856146e-06,
+      "loss": 0.9386,
+      "step": 170
+    },
+    {
+      "epoch": 0.8976377952755905,
+      "grad_norm": 0.45653306710128705,
+      "learning_rate": 4.0578310342365975e-06,
+      "loss": 0.9616,
+      "step": 171
+    },
+    {
+      "epoch": 0.9028871391076115,
+      "grad_norm": 0.3773630567359451,
+      "learning_rate": 4.031846755901785e-06,
+      "loss": 0.9285,
+      "step": 172
+    },
+    {
+      "epoch": 0.9081364829396326,
+      "grad_norm": 0.3644595191521506,
+      "learning_rate": 4.005774382006182e-06,
+      "loss": 0.9663,
+      "step": 173
+    },
+    {
+      "epoch": 0.9133858267716536,
+      "grad_norm": 0.3539767481135477,
+      "learning_rate": 3.97961613852025e-06,
+      "loss": 0.9564,
+      "step": 174
+    },
+    {
+      "epoch": 0.9186351706036745,
+      "grad_norm": 0.3819676152776953,
+      "learning_rate": 3.953374258745705e-06,
+      "loss": 0.9607,
+      "step": 175
+    },
+    {
+      "epoch": 0.9238845144356955,
+      "grad_norm": 0.38397675786726637,
+      "learning_rate": 3.927050983124842e-06,
+      "loss": 0.9539,
+      "step": 176
+    },
+    {
+      "epoch": 0.9291338582677166,
+      "grad_norm": 0.3979084367711538,
+      "learning_rate": 3.900648559049258e-06,
+      "loss": 0.9505,
+      "step": 177
+    },
+    {
+      "epoch": 0.9343832020997376,
+      "grad_norm": 0.3756154385935223,
+      "learning_rate": 3.874169240667974e-06,
+      "loss": 0.9519,
+      "step": 178
+    },
+    {
+      "epoch": 0.9396325459317585,
+      "grad_norm": 0.40551973597201274,
+      "learning_rate": 3.847615288694985e-06,
+      "loss": 0.9727,
+      "step": 179
+    },
+    {
+      "epoch": 0.9448818897637795,
+      "grad_norm": 0.4149625851710124,
+      "learning_rate": 3.820988970216249e-06,
+      "loss": 0.9464,
+      "step": 180
+    },
+    {
+      "epoch": 0.9501312335958005,
+      "grad_norm": 0.35739115830542967,
+      "learning_rate": 3.7942925584961272e-06,
+      "loss": 0.9427,
+      "step": 181
+    },
+    {
+      "epoch": 0.9553805774278216,
+      "grad_norm": 0.3759540038847051,
+      "learning_rate": 3.767528332783307e-06,
+      "loss": 0.9679,
+      "step": 182
+    },
+    {
+      "epoch": 0.9606299212598425,
+      "grad_norm": 0.3525867658299593,
+      "learning_rate": 3.740698578116199e-06,
+      "loss": 0.9183,
+      "step": 183
+    },
+    {
+      "epoch": 0.9658792650918635,
+      "grad_norm": 0.3557123352774738,
+      "learning_rate": 3.7138055851278564e-06,
+      "loss": 0.9383,
+      "step": 184
+    },
+    {
+      "epoch": 0.9711286089238845,
+      "grad_norm": 0.3623514252763418,
+      "learning_rate": 3.6868516498504025e-06,
+      "loss": 0.9246,
+      "step": 185
+    },
+    {
+      "epoch": 0.9763779527559056,
+      "grad_norm": 0.38495496418054853,
+      "learning_rate": 3.6598390735190066e-06,
+      "loss": 0.9612,
+      "step": 186
+    },
+    {
+      "epoch": 0.9816272965879265,
+      "grad_norm": 0.3648599004428126,
+      "learning_rate": 3.63277016237541e-06,
+      "loss": 0.9293,
+      "step": 187
+    },
+    {
+      "epoch": 0.9868766404199475,
+      "grad_norm": 0.38871547084803876,
+      "learning_rate": 3.6056472274710305e-06,
+      "loss": 0.9973,
+      "step": 188
+    },
+    {
+      "epoch": 0.9921259842519685,
+      "grad_norm": 0.38590844403642666,
+      "learning_rate": 3.578472584469651e-06,
+      "loss": 0.9457,
+      "step": 189
+    },
+    {
+      "epoch": 0.9973753280839895,
+      "grad_norm": 0.3872507088649178,
+      "learning_rate": 3.5512485534497116e-06,
+      "loss": 0.9462,
+      "step": 190
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.3872507088649178,
+      "learning_rate": 3.523977458706237e-06,
+      "loss": 0.9693,
+      "step": 191
+    },
+    {
+      "epoch": 1.005249343832021,
+      "grad_norm": 0.6232728744646114,
+      "learning_rate": 3.49666162855239e-06,
+      "loss": 0.887,
+      "step": 192
+    },
+    {
+      "epoch": 1.010498687664042,
+      "grad_norm": 0.4149641950734625,
+      "learning_rate": 3.469303395120693e-06,
+      "loss": 0.8826,
+      "step": 193
+    },
+    {
+      "epoch": 1.015748031496063,
+      "grad_norm": 0.37273340109017755,
+      "learning_rate": 3.441905094163913e-06,
+      "loss": 0.8893,
+      "step": 194
+    },
+    {
+      "epoch": 1.020997375328084,
+      "grad_norm": 0.4113832689982837,
+      "learning_rate": 3.414469064855647e-06,
+      "loss": 0.9205,
+      "step": 195
+    },
+    {
+      "epoch": 1.026246719160105,
+      "grad_norm": 0.49485155842511663,
+      "learning_rate": 3.3869976495906104e-06,
+      "loss": 0.9074,
+      "step": 196
+    },
+    {
+      "epoch": 1.031496062992126,
+      "grad_norm": 0.3736781934252868,
+      "learning_rate": 3.3594931937846498e-06,
+      "loss": 0.8966,
+      "step": 197
+    },
+    {
+      "epoch": 1.036745406824147,
+      "grad_norm": 0.3758650059773124,
+      "learning_rate": 3.3319580456745023e-06,
+      "loss": 0.8759,
+      "step": 198
+    },
+    {
+      "epoch": 1.041994750656168,
+      "grad_norm": 0.4056031624712629,
+      "learning_rate": 3.3043945561173092e-06,
+      "loss": 0.8788,
+      "step": 199
+    },
+    {
+      "epoch": 1.047244094488189,
+      "grad_norm": 0.36344982085137467,
+      "learning_rate": 3.2768050783899063e-06,
+      "loss": 0.873,
+      "step": 200
+    },
+    {
+      "epoch": 1.05249343832021,
+      "grad_norm": 0.3760103676246,
+      "learning_rate": 3.249191967987912e-06,
+      "loss": 0.899,
+      "step": 201
+    },
+    {
+      "epoch": 1.057742782152231,
+      "grad_norm": 0.39433477834527153,
+      "learning_rate": 3.221557582424622e-06,
+      "loss": 0.9019,
+      "step": 202
+    },
+    {
+      "epoch": 1.0629921259842519,
+      "grad_norm": 0.3595753440791428,
+      "learning_rate": 3.1939042810297328e-06,
+      "loss": 0.8781,
+      "step": 203
+    },
+    {
+      "epoch": 1.068241469816273,
+      "grad_norm": 0.3743448170598354,
+      "learning_rate": 3.16623442474791e-06,
+      "loss": 0.8689,
+      "step": 204
+    },
+    {
+      "epoch": 1.073490813648294,
+      "grad_norm": 0.3618551186966609,
+      "learning_rate": 3.138550375937219e-06,
+      "loss": 0.9094,
+      "step": 205
+    },
+    {
+      "epoch": 1.078740157480315,
+      "grad_norm": 0.36577516842050983,
+      "learning_rate": 3.1108544981674356e-06,
+      "loss": 0.8668,
+      "step": 206
+    },
+    {
+      "epoch": 1.083989501312336,
+      "grad_norm": 0.3985134455319658,
+      "learning_rate": 3.0831491560182495e-06,
+      "loss": 0.9016,
+      "step": 207
+    },
+    {
+      "epoch": 1.0892388451443569,
+      "grad_norm": 0.37808489525197075,
+      "learning_rate": 3.0554367148773897e-06,
+      "loss": 0.895,
+      "step": 208
+    },
+    {
+      "epoch": 1.094488188976378,
+      "grad_norm": 0.4112784941005797,
+      "learning_rate": 3.027719540738673e-06,
+      "loss": 0.859,
+      "step": 209
+    },
+    {
+      "epoch": 1.099737532808399,
+      "grad_norm": 0.3830296759827936,
+      "learning_rate": 3e-06,
+      "loss": 0.8569,
+      "step": 210
+    },
+    {
+      "epoch": 1.10498687664042,
+      "grad_norm": 0.3930755503999148,
+      "learning_rate": 2.972280459261328e-06,
+      "loss": 0.8774,
+      "step": 211
+    },
+    {
+      "epoch": 1.110236220472441,
+      "grad_norm": 0.36738851637178116,
+      "learning_rate": 2.944563285122611e-06,
+      "loss": 0.9086,
+      "step": 212
+    },
+    {
+      "epoch": 1.1154855643044619,
+      "grad_norm": 0.3897160841039193,
+      "learning_rate": 2.9168508439817515e-06,
+      "loss": 0.889,
+      "step": 213
+    },
+    {
+      "epoch": 1.120734908136483,
+      "grad_norm": 0.39858146379374537,
+      "learning_rate": 2.889145501832566e-06,
+      "loss": 0.8964,
+      "step": 214
+    },
+    {
+      "epoch": 1.125984251968504,
+      "grad_norm": 0.3739395525411432,
+      "learning_rate": 2.861449624062782e-06,
+      "loss": 0.8884,
+      "step": 215
+    },
+    {
+      "epoch": 1.1312335958005248,
+      "grad_norm": 0.3755768464864809,
+      "learning_rate": 2.83376557525209e-06,
+      "loss": 0.851,
+      "step": 216
+    },
+    {
+      "epoch": 1.136482939632546,
+      "grad_norm": 0.38260315757882735,
+      "learning_rate": 2.8060957189702674e-06,
+      "loss": 0.9152,
+      "step": 217
+    },
+    {
+      "epoch": 1.141732283464567,
+      "grad_norm": 0.4205379839527009,
+      "learning_rate": 2.7784424175753784e-06,
+      "loss": 0.8683,
+      "step": 218
+    },
+    {
+      "epoch": 1.1469816272965878,
+      "grad_norm": 0.38325260941818995,
+      "learning_rate": 2.7508080320120888e-06,
+      "loss": 0.8943,
+      "step": 219
+    },
+    {
+      "epoch": 1.152230971128609,
+      "grad_norm": 0.3763198826603672,
+      "learning_rate": 2.7231949216100943e-06,
+      "loss": 0.8676,
+      "step": 220
+    },
+    {
+      "epoch": 1.1574803149606299,
+      "grad_norm": 0.3767162287387105,
+      "learning_rate": 2.6956054438826918e-06,
+      "loss": 0.8482,
+      "step": 221
+    },
+    {
+      "epoch": 1.162729658792651,
+      "grad_norm": 0.3486273740901837,
+      "learning_rate": 2.668041954325498e-06,
+      "loss": 0.8879,
+      "step": 222
+    },
+    {
+      "epoch": 1.167979002624672,
+      "grad_norm": 0.39084218665366566,
+      "learning_rate": 2.640506806215351e-06,
+      "loss": 0.8679,
+      "step": 223
+    },
+    {
+      "epoch": 1.1732283464566928,
+      "grad_norm": 0.3538552501730603,
+      "learning_rate": 2.613002350409391e-06,
+      "loss": 0.8871,
+      "step": 224
+    },
+    {
+      "epoch": 1.178477690288714,
+      "grad_norm": 0.36544200913577,
+      "learning_rate": 2.585530935144354e-06,
+      "loss": 0.8616,
+      "step": 225
+    },
+    {
+      "epoch": 1.1837270341207349,
+      "grad_norm": 0.3985990462573467,
+      "learning_rate": 2.558094905836087e-06,
+      "loss": 0.8917,
+      "step": 226
+    },
+    {
+      "epoch": 1.188976377952756,
+      "grad_norm": 0.42608518999556655,
+      "learning_rate": 2.5306966048793067e-06,
+      "loss": 0.8817,
+      "step": 227
+    },
+    {
+      "epoch": 1.194225721784777,
+      "grad_norm": 0.37952769789031354,
+      "learning_rate": 2.5033383714476097e-06,
+      "loss": 0.8985,
+      "step": 228
+    },
+    {
+      "epoch": 1.1994750656167978,
+      "grad_norm": 0.40804864076806885,
+      "learning_rate": 2.4760225412937633e-06,
+      "loss": 0.9073,
+      "step": 229
+    },
+    {
+      "epoch": 1.204724409448819,
+      "grad_norm": 0.4167713152946991,
+      "learning_rate": 2.4487514465502885e-06,
+      "loss": 0.8566,
+      "step": 230
+    },
+    {
+      "epoch": 1.20997375328084,
+      "grad_norm": 0.4022153540631621,
+      "learning_rate": 2.42152741553035e-06,
+      "loss": 0.8713,
+      "step": 231
+    },
+    {
+      "epoch": 1.2152230971128608,
+      "grad_norm": 0.4222065137992956,
+      "learning_rate": 2.39435277252897e-06,
+      "loss": 0.9035,
+      "step": 232
+    },
+    {
+      "epoch": 1.220472440944882,
+      "grad_norm": 0.3666365807384159,
+      "learning_rate": 2.3672298376245908e-06,
+      "loss": 0.8637,
+      "step": 233
+    },
+    {
+      "epoch": 1.2257217847769029,
+      "grad_norm": 0.3976853335036615,
+      "learning_rate": 2.3401609264809953e-06,
+      "loss": 0.9398,
+      "step": 234
+    },
+    {
+      "epoch": 1.2309711286089238,
+      "grad_norm": 0.37956934109451046,
+      "learning_rate": 2.3131483501495985e-06,
+      "loss": 0.8353,
+      "step": 235
+    },
+    {
+      "epoch": 1.236220472440945,
+      "grad_norm": 0.33722056538083744,
+      "learning_rate": 2.2861944148721446e-06,
+      "loss": 0.8786,
+      "step": 236
+    },
+    {
+      "epoch": 1.2414698162729658,
+      "grad_norm": 0.49777382093647954,
+      "learning_rate": 2.2593014218838e-06,
+      "loss": 0.8834,
+      "step": 237
+    },
+    {
+      "epoch": 1.246719160104987,
+      "grad_norm": 0.35315516410389436,
+      "learning_rate": 2.232471667216693e-06,
+      "loss": 0.8442,
+      "step": 238
+    },
+    {
+      "epoch": 1.2519685039370079,
+      "grad_norm": 0.3816124424363711,
+      "learning_rate": 2.2057074415038725e-06,
+      "loss": 0.8573,
+      "step": 239
+    },
+    {
+      "epoch": 1.257217847769029,
+      "grad_norm": 0.36319142999803095,
+      "learning_rate": 2.1790110297837514e-06,
+      "loss": 0.8481,
+      "step": 240
+    },
+    {
+      "epoch": 1.26246719160105,
+      "grad_norm": 0.34672889281207053,
+      "learning_rate": 2.152384711305015e-06,
+      "loss": 0.8623,
+      "step": 241
+    },
+    {
+      "epoch": 1.2677165354330708,
+      "grad_norm": 0.37448151544392105,
+      "learning_rate": 2.1258307593320262e-06,
+      "loss": 0.8751,
+      "step": 242
+    },
+    {
+      "epoch": 1.272965879265092,
+      "grad_norm": 0.37082567424502005,
+      "learning_rate": 2.099351440950742e-06,
+      "loss": 0.8914,
+      "step": 243
+    },
+    {
+      "epoch": 1.2782152230971129,
+      "grad_norm": 0.39074992783073415,
+      "learning_rate": 2.072949016875158e-06,
+      "loss": 0.9222,
+      "step": 244
+    },
+    {
+      "epoch": 1.2834645669291338,
+      "grad_norm": 0.4150437401629804,
+      "learning_rate": 2.046625741254295e-06,
+      "loss": 0.9475,
+      "step": 245
+    },
+    {
+      "epoch": 1.288713910761155,
+      "grad_norm": 0.4504166670407193,
+      "learning_rate": 2.0203838614797505e-06,
+      "loss": 0.9026,
+      "step": 246
+    },
+    {
+      "epoch": 1.2939632545931758,
+      "grad_norm": 0.38345958484903814,
+      "learning_rate": 1.994225617993819e-06,
+      "loss": 0.9074,
+      "step": 247
+    },
+    {
+      "epoch": 1.2992125984251968,
+      "grad_norm": 0.37086048031752866,
+      "learning_rate": 1.9681532440982154e-06,
+      "loss": 0.8755,
+      "step": 248
+    },
+    {
+      "epoch": 1.304461942257218,
+      "grad_norm": 0.3775524407980251,
+      "learning_rate": 1.942168965763402e-06,
+      "loss": 0.8986,
+      "step": 249
+    },
+    {
+      "epoch": 1.3097112860892388,
+      "grad_norm": 0.364796377340789,
+      "learning_rate": 1.916275001438541e-06,
+      "loss": 0.867,
+      "step": 250
+    },
+    {
+      "epoch": 1.3149606299212597,
+      "grad_norm": 0.3705604843330414,
+      "learning_rate": 1.8904735618620928e-06,
+      "loss": 0.8875,
+      "step": 251
+    },
+    {
+      "epoch": 1.3202099737532809,
+      "grad_norm": 0.3847344001283667,
+      "learning_rate": 1.8647668498730693e-06,
+      "loss": 0.8678,
+      "step": 252
+    },
+    {
+      "epoch": 1.3254593175853018,
+      "grad_norm": 0.3507183610862785,
+      "learning_rate": 1.8391570602229647e-06,
+      "loss": 0.8895,
+      "step": 253
+    },
+    {
+      "epoch": 1.330708661417323,
+      "grad_norm": 0.34464955572346173,
+      "learning_rate": 1.8136463793883725e-06,
+      "loss": 0.9112,
+      "step": 254
+    },
+    {
+      "epoch": 1.3359580052493438,
+      "grad_norm": 0.3804540728076062,
+      "learning_rate": 1.7882369853843155e-06,
+      "loss": 0.8818,
+      "step": 255
+    },
+    {
+      "epoch": 1.341207349081365,
+      "grad_norm": 0.38671544491057547,
+      "learning_rate": 1.76293104757829e-06,
+      "loss": 0.8712,
+      "step": 256
+    },
+    {
+      "epoch": 1.3464566929133859,
+      "grad_norm": 0.35028636565033566,
+      "learning_rate": 1.7377307265050559e-06,
+      "loss": 0.8795,
+      "step": 257
+    },
+    {
+      "epoch": 1.3517060367454068,
+      "grad_norm": 0.3596694021401425,
+      "learning_rate": 1.7126381736821732e-06,
+      "loss": 0.8791,
+      "step": 258
+    },
+    {
+      "epoch": 1.356955380577428,
+      "grad_norm": 0.3833574983214166,
+      "learning_rate": 1.6876555314263213e-06,
+      "loss": 0.9108,
+      "step": 259
+    },
+    {
+      "epoch": 1.3622047244094488,
+      "grad_norm": 0.3701840047085969,
+      "learning_rate": 1.6627849326703855e-06,
+      "loss": 0.8695,
+      "step": 260
+    },
+    {
+      "epoch": 1.3674540682414698,
+      "grad_norm": 0.36098816535443995,
+      "learning_rate": 1.6380285007813598e-06,
+      "loss": 0.876,
+      "step": 261
+    },
+    {
+      "epoch": 1.372703412073491,
+      "grad_norm": 0.3900890284585014,
+      "learning_rate": 1.6133883493790609e-06,
+      "loss": 0.8498,
+      "step": 262
+    },
+    {
+      "epoch": 1.3779527559055118,
+      "grad_norm": 0.34906551126755136,
+      "learning_rate": 1.5888665821556724e-06,
+      "loss": 0.8513,
+      "step": 263
+    },
+    {
+      "epoch": 1.3832020997375327,
+      "grad_norm": 0.3753732283477496,
+      "learning_rate": 1.5644652926961407e-06,
+      "loss": 0.8714,
+      "step": 264
+    },
+    {
+      "epoch": 1.3884514435695539,
+      "grad_norm": 0.34748864593560347,
+      "learning_rate": 1.5401865642994315e-06,
+      "loss": 0.9124,
+      "step": 265
+    },
+    {
+      "epoch": 1.3937007874015748,
+      "grad_norm": 0.36698053817770165,
+      "learning_rate": 1.5160324698006642e-06,
+      "loss": 0.8814,
+      "step": 266
+    },
+    {
+      "epoch": 1.3989501312335957,
+      "grad_norm": 0.4000964153653425,
+      "learning_rate": 1.4920050713941398e-06,
+      "loss": 0.9082,
+      "step": 267
+    },
+    {
+      "epoch": 1.4041994750656168,
+      "grad_norm": 0.3985391177875817,
+      "learning_rate": 1.4681064204572798e-06,
+      "loss": 0.8749,
+      "step": 268
+    },
+    {
+      "epoch": 1.4094488188976377,
+      "grad_norm": 0.3578122677174226,
+      "learning_rate": 1.4443385573754837e-06,
+      "loss": 0.8608,
+      "step": 269
+    },
+    {
+      "epoch": 1.4146981627296589,
+      "grad_norm": 0.3576093239254431,
+      "learning_rate": 1.4207035113679322e-06,
+      "loss": 0.8798,
+      "step": 270
+    },
+    {
+      "epoch": 1.4199475065616798,
+      "grad_norm": 0.35299639204379674,
+      "learning_rate": 1.3972033003143348e-06,
+      "loss": 0.8972,
+      "step": 271
+    },
+    {
+      "epoch": 1.425196850393701,
+      "grad_norm": 0.3937775289907907,
+      "learning_rate": 1.3738399305826516e-06,
+      "loss": 0.8736,
+      "step": 272
+    },
+    {
+      "epoch": 1.4304461942257218,
+      "grad_norm": 0.3691998032129419,
+      "learning_rate": 1.3506153968577983e-06,
+      "loss": 0.8667,
+      "step": 273
+    },
+    {
+      "epoch": 1.4356955380577427,
+      "grad_norm": 0.35764876894907843,
+      "learning_rate": 1.3275316819713435e-06,
+      "loss": 0.882,
+      "step": 274
+    },
+    {
+      "epoch": 1.4409448818897639,
+      "grad_norm": 0.3859579688778526,
+      "learning_rate": 1.3045907567322243e-06,
+      "loss": 0.844,
+      "step": 275
+    },
+    {
+      "epoch": 1.4461942257217848,
+      "grad_norm": 0.3736621084680505,
+      "learning_rate": 1.2817945797584844e-06,
+      "loss": 0.8525,
+      "step": 276
+    },
+    {
+      "epoch": 1.4514435695538057,
+      "grad_norm": 0.36602372507940695,
+      "learning_rate": 1.2591450973100532e-06,
+      "loss": 0.8577,
+      "step": 277
+    },
+    {
+      "epoch": 1.4566929133858268,
+      "grad_norm": 0.37926054124030645,
+      "learning_rate": 1.236644243122581e-06,
+      "loss": 0.8837,
+      "step": 278
+    },
+    {
+      "epoch": 1.4619422572178478,
+      "grad_norm": 0.3680022216795608,
+      "learning_rate": 1.214293938242344e-06,
+      "loss": 0.8984,
+      "step": 279
+    },
+    {
+      "epoch": 1.4671916010498687,
+      "grad_norm": 0.37824901927870175,
+      "learning_rate": 1.1920960908622313e-06,
+      "loss": 0.8745,
+      "step": 280
+    },
+    {
+      "epoch": 1.4724409448818898,
+      "grad_norm": 0.3489273490529577,
+      "learning_rate": 1.17005259615883e-06,
+      "loss": 0.8628,
+      "step": 281
+    },
+    {
+      "epoch": 1.4776902887139107,
+      "grad_norm": 0.3735770062938505,
+      "learning_rate": 1.1481653361306215e-06,
+      "loss": 0.8619,
+      "step": 282
+    },
+    {
+      "epoch": 1.4829396325459316,
+      "grad_norm": 0.3458041443504503,
+      "learning_rate": 1.1264361794373032e-06,
+      "loss": 0.8761,
+      "step": 283
+    },
+    {
+      "epoch": 1.4881889763779528,
+      "grad_norm": 0.35998420937846626,
+      "learning_rate": 1.104866981240248e-06,
+      "loss": 0.8844,
+      "step": 284
+    },
+    {
+      "epoch": 1.4934383202099737,
+      "grad_norm": 0.4029178073367971,
+      "learning_rate": 1.0834595830441168e-06,
+      "loss": 0.8511,
+      "step": 285
+    },
+    {
+      "epoch": 1.4986876640419948,
+      "grad_norm": 0.3763622650814437,
+      "learning_rate": 1.0622158125396387e-06,
+      "loss": 0.8599,
+      "step": 286
+    },
+    {
+      "epoch": 1.5039370078740157,
+      "grad_norm": 0.3845965137728459,
+      "learning_rate": 1.0411374834475678e-06,
+      "loss": 0.9062,
+      "step": 287
+    },
+    {
+      "epoch": 1.5091863517060369,
+      "grad_norm": 0.34964825506869784,
+      "learning_rate": 1.020226395363833e-06,
+      "loss": 0.8525,
+      "step": 288
+    },
+    {
+      "epoch": 1.5144356955380578,
+      "grad_norm": 0.38214019455395715,
+      "learning_rate": 9.994843336058968e-07,
+      "loss": 0.8686,
+      "step": 289
+    },
+    {
+      "epoch": 1.5196850393700787,
+      "grad_norm": 0.3808975526218143,
+      "learning_rate": 9.789130690603284e-07,
+      "loss": 0.8537,
+      "step": 290
+    },
+    {
+      "epoch": 1.5249343832020998,
+      "grad_norm": 0.3761982373529746,
+      "learning_rate": 9.585143580316153e-07,
+      "loss": 0.8826,
+      "step": 291
+    },
+    {
+      "epoch": 1.5301837270341208,
+      "grad_norm": 0.351389916026518,
+      "learning_rate": 9.382899420922119e-07,
+      "loss": 0.8683,
+      "step": 292
+    },
+    {
+      "epoch": 1.5354330708661417,
+      "grad_norm": 0.3711139029247798,
+      "learning_rate": 9.182415479338512e-07,
+      "loss": 0.8878,
+      "step": 293
+    },
+    {
+      "epoch": 1.5406824146981628,
+      "grad_norm": 0.3717732780588312,
+      "learning_rate": 8.983708872201271e-07,
+      "loss": 0.8585,
+      "step": 294
+    },
+    {
+      "epoch": 1.5459317585301837,
+      "grad_norm": 0.3742271193984993,
+      "learning_rate": 8.786796564403577e-07,
+      "loss": 0.8579,
+      "step": 295
+    },
+    {
+      "epoch": 1.5511811023622046,
+      "grad_norm": 0.42523434659053005,
+      "learning_rate": 8.591695367647433e-07,
+      "loss": 0.8746,
+      "step": 296
+    },
+    {
+      "epoch": 1.5564304461942258,
+      "grad_norm": 0.3794388162880317,
+      "learning_rate": 8.398421939008367e-07,
+      "loss": 0.8479,
+      "step": 297
+    },
+    {
+      "epoch": 1.5616797900262467,
+      "grad_norm": 0.3588910082794427,
+      "learning_rate": 8.206992779513281e-07,
+      "loss": 0.8329,
+      "step": 298
+    },
+    {
+      "epoch": 1.5669291338582676,
+      "grad_norm": 0.37000389491476643,
+      "learning_rate": 8.017424232731664e-07,
+      "loss": 0.8693,
+      "step": 299
+    },
+    {
+      "epoch": 1.5721784776902887,
+      "grad_norm": 0.4003207798760719,
+      "learning_rate": 7.829732483380231e-07,
+      "loss": 0.8886,
+      "step": 300
+    },
+    {
+      "epoch": 1.5774278215223099,
+      "grad_norm": 0.37170546863230536,
+      "learning_rate": 7.64393355594112e-07,
+      "loss": 0.9035,
+      "step": 301
+    },
+    {
+      "epoch": 1.5826771653543306,
+      "grad_norm": 0.35766484669954807,
+      "learning_rate": 7.4600433132938e-07,
+      "loss": 0.848,
+      "step": 302
+    },
+    {
+      "epoch": 1.5879265091863517,
+      "grad_norm": 0.33788894124632585,
+      "learning_rate": 7.278077455360717e-07,
+      "loss": 0.8545,
+      "step": 303
+    },
+    {
+      "epoch": 1.5931758530183728,
+      "grad_norm": 0.36604275227388566,
+      "learning_rate": 7.09805151776691e-07,
+      "loss": 0.8415,
+      "step": 304
+    },
+    {
+      "epoch": 1.5984251968503937,
+      "grad_norm": 0.33845675028801603,
+      "learning_rate": 6.919980870513633e-07,
+      "loss": 0.8472,
+      "step": 305
+    },
+    {
+      "epoch": 1.6036745406824147,
+      "grad_norm": 0.35701657873038517,
+      "learning_rate": 6.743880716666104e-07,
+      "loss": 0.8496,
+      "step": 306
+    },
+    {
+      "epoch": 1.6089238845144358,
+      "grad_norm": 0.3779107207471187,
+      "learning_rate": 6.569766091055539e-07,
+      "loss": 0.8495,
+      "step": 307
+    },
+    {
+      "epoch": 1.6141732283464567,
+      "grad_norm": 0.3872432739805792,
+      "learning_rate": 6.397651858995504e-07,
+      "loss": 0.851,
+      "step": 308
+    },
+    {
+      "epoch": 1.6194225721784776,
+      "grad_norm": 0.3595059488828886,
+      "learning_rate": 6.227552715012785e-07,
+      "loss": 0.8855,
+      "step": 309
+    },
+    {
+      "epoch": 1.6246719160104988,
+      "grad_norm": 0.37515207797149636,
+      "learning_rate": 6.059483181592815e-07,
+      "loss": 0.8858,
+      "step": 310
+    },
+    {
+      "epoch": 1.6299212598425197,
+      "grad_norm": 0.3629206169447269,
+      "learning_rate": 5.893457607939788e-07,
+      "loss": 0.8807,
+      "step": 311
+    },
+    {
+      "epoch": 1.6351706036745406,
+      "grad_norm": 0.3545486266254271,
+      "learning_rate": 5.72949016875158e-07,
+      "loss": 0.8955,
+      "step": 312
+    },
+    {
+      "epoch": 1.6404199475065617,
+      "grad_norm": 0.3432909080087375,
+      "learning_rate": 5.56759486300955e-07,
+      "loss": 0.8681,
+      "step": 313
+    },
+    {
+      "epoch": 1.6456692913385826,
+      "grad_norm": 0.3639385301913205,
+      "learning_rate": 5.40778551278337e-07,
+      "loss": 0.8733,
+      "step": 314
+    },
+    {
+      "epoch": 1.6509186351706036,
+      "grad_norm": 0.38920489291790045,
+      "learning_rate": 5.250075762050935e-07,
+      "loss": 0.8745,
+      "step": 315
+    },
+    {
+      "epoch": 1.6561679790026247,
+      "grad_norm": 0.3618641411341515,
+      "learning_rate": 5.094479075533486e-07,
+      "loss": 0.8832,
+      "step": 316
+    },
+    {
+      "epoch": 1.6614173228346458,
+      "grad_norm": 0.38194003303963936,
+      "learning_rate": 4.941008737546039e-07,
+      "loss": 0.882,
+      "step": 317
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.36004508832846943,
+      "learning_rate": 4.789677850863222e-07,
+      "loss": 0.8754,
+      "step": 318
+    },
+    {
+      "epoch": 1.6719160104986877,
+      "grad_norm": 0.3713147294795857,
+      "learning_rate": 4.6404993356005967e-07,
+      "loss": 0.8496,
+      "step": 319
+    },
+    {
+      "epoch": 1.6771653543307088,
+      "grad_norm": 0.40332844918244803,
+      "learning_rate": 4.4934859281115804e-07,
+      "loss": 0.8985,
+      "step": 320
+    },
+    {
+      "epoch": 1.6824146981627297,
+      "grad_norm": 0.37460256078858306,
+      "learning_rate": 4.34865017990007e-07,
+      "loss": 0.8997,
+      "step": 321
+    },
+    {
+      "epoch": 1.6876640419947506,
+      "grad_norm": 0.34681859755129757,
+      "learning_rate": 4.2060044565488264e-07,
+      "loss": 0.8596,
+      "step": 322
+    },
+    {
+      "epoch": 1.6929133858267718,
+      "grad_norm": 0.36407790059697526,
+      "learning_rate": 4.0655609366637635e-07,
+      "loss": 0.8891,
+      "step": 323
+    },
+    {
+      "epoch": 1.6981627296587927,
+      "grad_norm": 0.38031424057937346,
+      "learning_rate": 3.9273316108341493e-07,
+      "loss": 0.9026,
+      "step": 324
+    },
+    {
+      "epoch": 1.7034120734908136,
+      "grad_norm": 0.37078158732866173,
+      "learning_rate": 3.791328280608916e-07,
+      "loss": 0.8676,
+      "step": 325
+    },
+    {
+      "epoch": 1.7086614173228347,
+      "grad_norm": 0.3744777105843697,
+      "learning_rate": 3.657562557489063e-07,
+      "loss": 0.8692,
+      "step": 326
+    },
+    {
+      "epoch": 1.7139107611548556,
+      "grad_norm": 0.35275476224983093,
+      "learning_rate": 3.52604586193633e-07,
+      "loss": 0.878,
+      "step": 327
+    },
+    {
+      "epoch": 1.7191601049868765,
+      "grad_norm": 0.3845721220969486,
+      "learning_rate": 3.396789422398143e-07,
+      "loss": 0.8715,
+      "step": 328
+    },
+    {
+      "epoch": 1.7244094488188977,
+      "grad_norm": 0.38567935425060995,
+      "learning_rate": 3.269804274348966e-07,
+      "loss": 0.8552,
+      "step": 329
+    },
+    {
+      "epoch": 1.7296587926509186,
+      "grad_norm": 0.36984914245968326,
+      "learning_rate": 3.145101259348133e-07,
+      "loss": 0.8905,
+      "step": 330
+    },
+    {
+      "epoch": 1.7349081364829395,
+      "grad_norm": 0.3862126346234947,
+      "learning_rate": 3.022691024114234e-07,
+      "loss": 0.8759,
+      "step": 331
+    },
+    {
+      "epoch": 1.7401574803149606,
+      "grad_norm": 0.37276644156643624,
+      "learning_rate": 2.9025840196161345e-07,
+      "loss": 0.8996,
+      "step": 332
+    },
+    {
+      "epoch": 1.7454068241469818,
+      "grad_norm": 0.3619714746103851,
+      "learning_rate": 2.784790500180704e-07,
+      "loss": 0.8734,
+      "step": 333
+    },
+    {
+      "epoch": 1.7506561679790025,
+      "grad_norm": 0.3489848842196673,
+      "learning_rate": 2.6693205226173466e-07,
+      "loss": 0.852,
+      "step": 334
+    },
+    {
+      "epoch": 1.7559055118110236,
+      "grad_norm": 0.47292345630417715,
+      "learning_rate": 2.556183945359369e-07,
+      "loss": 0.8524,
+      "step": 335
+    },
+    {
+      "epoch": 1.7611548556430447,
+      "grad_norm": 0.3454751472880757,
+      "learning_rate": 2.4453904276223093e-07,
+      "loss": 0.8639,
+      "step": 336
+    },
+    {
+      "epoch": 1.7664041994750657,
+      "grad_norm": 0.39520326916346893,
+      "learning_rate": 2.3369494285792648e-07,
+      "loss": 0.9011,
+      "step": 337
+    },
+    {
+      "epoch": 1.7716535433070866,
+      "grad_norm": 0.36215879824858,
+      "learning_rate": 2.2308702065533138e-07,
+      "loss": 0.8475,
+      "step": 338
+    },
+    {
+      "epoch": 1.7769028871391077,
+      "grad_norm": 0.3785025816595213,
+      "learning_rate": 2.1271618182270402e-07,
+      "loss": 0.8785,
+      "step": 339
+    },
+    {
+      "epoch": 1.7821522309711286,
+      "grad_norm": 0.35017564001831825,
+      "learning_rate": 2.0258331178693291e-07,
+      "loss": 0.9251,
+      "step": 340
+    },
+    {
+      "epoch": 1.7874015748031495,
+      "grad_norm": 0.3736191583691154,
+      "learning_rate": 1.926892756579427e-07,
+      "loss": 0.8638,
+      "step": 341
+    },
+    {
+      "epoch": 1.7926509186351707,
+      "grad_norm": 0.3542533426451256,
+      "learning_rate": 1.8303491815483076e-07,
+      "loss": 0.8501,
+      "step": 342
+    },
+    {
+      "epoch": 1.7979002624671916,
+      "grad_norm": 0.3653513060765524,
+      "learning_rate": 1.7362106353375107e-07,
+      "loss": 0.8704,
+      "step": 343
+    },
+    {
+      "epoch": 1.8031496062992125,
+      "grad_norm": 0.3699565968914539,
+      "learning_rate": 1.6444851551754158e-07,
+      "loss": 0.8659,
+      "step": 344
+    },
+    {
+      "epoch": 1.8083989501312336,
+      "grad_norm": 0.357867752615946,
+      "learning_rate": 1.5551805722710245e-07,
+      "loss": 0.8802,
+      "step": 345
+    },
+    {
+      "epoch": 1.8136482939632546,
+      "grad_norm": 0.37125992932731333,
+      "learning_rate": 1.4683045111453941e-07,
+      "loss": 0.8368,
+      "step": 346
+    },
+    {
+      "epoch": 1.8188976377952755,
+      "grad_norm": 0.3685836112400432,
+      "learning_rate": 1.3838643889806568e-07,
+      "loss": 0.9235,
+      "step": 347
+    },
+    {
+      "epoch": 1.8241469816272966,
+      "grad_norm": 0.36408877164595227,
+      "learning_rate": 1.3018674149867782e-07,
+      "loss": 0.8799,
+      "step": 348
+    },
+    {
+      "epoch": 1.8293963254593177,
+      "grad_norm": 0.3642144586766023,
+      "learning_rate": 1.2223205897860533e-07,
+      "loss": 0.8777,
+      "step": 349
+    },
+    {
+      "epoch": 1.8346456692913384,
+      "grad_norm": 0.39073442531405206,
+      "learning_rate": 1.1452307048154286e-07,
+      "loss": 0.8797,
+      "step": 350
+    },
+    {
+      "epoch": 1.8398950131233596,
+      "grad_norm": 0.3810371400467593,
+      "learning_rate": 1.0706043417466549e-07,
+      "loss": 0.91,
+      "step": 351
+    },
+    {
+      "epoch": 1.8451443569553807,
+      "grad_norm": 0.35959663657652996,
+      "learning_rate": 9.984478719243772e-08,
+      "loss": 0.8714,
+      "step": 352
+    },
+    {
+      "epoch": 1.8503937007874016,
+      "grad_norm": 0.36819474514599226,
+      "learning_rate": 9.287674558221737e-08,
+      "loss": 0.9048,
+      "step": 353
+    },
+    {
+      "epoch": 1.8556430446194225,
+      "grad_norm": 0.3492451150451855,
+      "learning_rate": 8.615690425165823e-08,
+      "loss": 0.8589,
+      "step": 354
+    },
+    {
+      "epoch": 1.8608923884514437,
+      "grad_norm": 0.3692991236824256,
+      "learning_rate": 7.968583691792142e-08,
+      "loss": 0.8502,
+      "step": 355
+    },
+    {
+      "epoch": 1.8661417322834646,
+      "grad_norm": 0.36506034919430097,
+      "learning_rate": 7.34640960586902e-08,
+      "loss": 0.8948,
+      "step": 356
+    },
+    {
+      "epoch": 1.8713910761154855,
+      "grad_norm": 0.39656772129331486,
+      "learning_rate": 6.749221286500273e-08,
+      "loss": 0.872,
+      "step": 357
+    },
+    {
+      "epoch": 1.8766404199475066,
+      "grad_norm": 0.3766497135471153,
+      "learning_rate": 6.177069719590234e-08,
+      "loss": 0.8459,
+      "step": 358
+    },
+    {
+      "epoch": 1.8818897637795275,
+      "grad_norm": 0.3420155942064905,
+      "learning_rate": 5.6300037534904644e-08,
+      "loss": 0.8797,
+      "step": 359
+    },
+    {
+      "epoch": 1.8871391076115485,
+      "grad_norm": 0.3790798456525195,
+      "learning_rate": 5.108070094829465e-08,
+      "loss": 0.8374,
+      "step": 360
+    },
+    {
+      "epoch": 1.8923884514435696,
+      "grad_norm": 0.36574760138577367,
+      "learning_rate": 4.6113133045249225e-08,
+      "loss": 0.8507,
+      "step": 361
+    },
+    {
+      "epoch": 1.8976377952755905,
+      "grad_norm": 0.39215283711659366,
+      "learning_rate": 4.139775793979228e-08,
+      "loss": 0.9416,
+      "step": 362
+    },
+    {
+      "epoch": 1.9028871391076114,
+      "grad_norm": 0.33743123097312766,
+      "learning_rate": 3.693497821458702e-08,
+      "loss": 0.8469,
+      "step": 363
+    },
+    {
+      "epoch": 1.9081364829396326,
+      "grad_norm": 0.365685642660726,
+      "learning_rate": 3.272517488656213e-08,
+      "loss": 0.8809,
+      "step": 364
+    },
+    {
+      "epoch": 1.9133858267716537,
+      "grad_norm": 0.35701737992686006,
+      "learning_rate": 2.876870737438475e-08,
+      "loss": 0.8576,
+      "step": 365
+    },
+    {
+      "epoch": 1.9186351706036744,
+      "grad_norm": 0.4717726718121241,
+      "learning_rate": 2.506591346777176e-08,
+      "loss": 0.8882,
+      "step": 366
+    },
+    {
+      "epoch": 1.9238845144356955,
+      "grad_norm": 0.3644123714802389,
+      "learning_rate": 2.1617109298653126e-08,
+      "loss": 0.8806,
+      "step": 367
+    },
+    {
+      "epoch": 1.9291338582677167,
+      "grad_norm": 0.37039268056716995,
+      "learning_rate": 1.842258931417917e-08,
+      "loss": 0.8646,
+      "step": 368
+    },
+    {
+      "epoch": 1.9343832020997376,
+      "grad_norm": 0.38307646079911417,
+      "learning_rate": 1.5482626251583364e-08,
+      "loss": 0.8605,
+      "step": 369
+    },
+    {
+      "epoch": 1.9396325459317585,
+      "grad_norm": 0.34939756521186505,
+      "learning_rate": 1.2797471114896598e-08,
+      "loss": 0.8605,
+      "step": 370
+    },
+    {
+      "epoch": 1.9448818897637796,
+      "grad_norm": 0.3630644212614912,
+      "learning_rate": 1.0367353153516335e-08,
+      "loss": 0.8874,
+      "step": 371
+    },
+    {
+      "epoch": 1.9501312335958005,
+      "grad_norm": 0.36486914850748925,
+      "learning_rate": 8.192479842635937e-09,
+      "loss": 0.8488,
+      "step": 372
+    },
+    {
+      "epoch": 1.9553805774278215,
+      "grad_norm": 0.3477602485063963,
+      "learning_rate": 6.273036865529158e-09,
+      "loss": 0.8865,
+      "step": 373
+    },
+    {
+      "epoch": 1.9606299212598426,
+      "grad_norm": 0.3699792979427579,
+      "learning_rate": 4.6091880976981695e-09,
+      "loss": 0.8552,
+      "step": 374
+    },
+    {
+      "epoch": 1.9658792650918635,
+      "grad_norm": 0.36569117211522567,
+      "learning_rate": 3.201075592882741e-09,
+      "loss": 0.8771,
+      "step": 375
+    },
+    {
+      "epoch": 1.9711286089238844,
+      "grad_norm": 0.3673522439384638,
+      "learning_rate": 2.0488195709312816e-09,
+      "loss": 0.8316,
+      "step": 376
+    },
+    {
+      "epoch": 1.9763779527559056,
+      "grad_norm": 0.40741119157495514,
+      "learning_rate": 1.152518407537717e-09,
+      "loss": 0.8686,
+      "step": 377
+    },
+    {
+      "epoch": 1.9816272965879265,
+      "grad_norm": 0.3922247762663656,
+      "learning_rate": 5.122486258418713e-10,
+      "loss": 0.88,
+      "step": 378
+    },
+    {
+      "epoch": 1.9868766404199474,
+      "grad_norm": 0.36724366518491103,
+      "learning_rate": 1.2806488989636211e-10,
+      "loss": 0.863,
+      "step": 379
+    },
+    {
+      "epoch": 1.9921259842519685,
+      "grad_norm": 0.37347664771694333,
+      "learning_rate": 0.0,
+      "loss": 0.8553,
+      "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:b860c7e838727b1a9d8001f5c7a769bd0d63566ea45620719245b6beb59f1cd9
+size 8376

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)