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 sha256:9cf10ce50a8d003eaff8d51daae1bc78a87469963f79d7c53a3ce107db3666e5
+size 4891730992

model-00002-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4e81bf27bad3b96697389808a44a571dbff4db739f57f41582ec0f7db527d76
+size 4876059352

model-00003-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcf6bf13b7617ff05293691beafd48c4a928a91fba50e1e1388a4a5c3d269109
+size 4876059384

model-00004-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ac13aaae6bdd992d8a58d0e596db8d562469b45316b604efb60b7c3780ccf8b
+size 4876059416

model-00005-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b0c3632f86a8031f6cd02e8b6f7813b235993f6a7a65ca7ca3529885e4b24f0
+size 4876059416

model-00006-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01b5ffdf2e2c4c78ba4a60e3bc3014ee88c48a6069ecf274b06882eb679af1db
+size 4876059416

model-00007-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a48c262fd09b64848047caca69fc42deb1c0ef7d5f059f3163b708128737319a
+size 4876059416

model-00008-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8d2a6506e468823609ad1a0c20c9b137c1fb686cc179c052bd0a576340e433e
+size 4876059416

model-00009-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aeb06318c94ab8ae043d59a01f8b66b39df677db6bef203523608b396d823fe
+size 4876059416

model-00010-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec1eed2a46b3a63174b1020cf13f4b3841e6a7d0fb1e28ead346fb0d36a68f60
+size 4876059416

model-00011-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c0cdd742a90b658ebd2e3ba8c13768db9306f507e2615ae25f9a29c38fbf401d
+size 4876059416

model-00012-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82e664c4014f96f9ce810b8804b2a30aa18f9896b4c2760974f39e0e8a1c94cf
+size 4876059416

model-00013-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77bae41788091207a2adafa153a080d989fc7dec91f121873d81b4ed1ba455e9
+size 4876059416

model-00014-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ece41c61d65c36120db9a4efe3e9602ecf07c02eacb1462298a677d1e20477f4
+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 sha256:719711f95f91b896df7e7013497079a4d63ea6ed67bafa499f5687f091965ffa
+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": 1.134754623075341,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 1.1087,
+      "step": 1
+    },
+    {
+      "epoch": 0.010498687664041995,
+      "grad_norm": 1.1234145683168772,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.1356,
+      "step": 2
+    },
+    {
+      "epoch": 0.015748031496062992,
+      "grad_norm": 1.0799860590372758,
+      "learning_rate": 3e-06,
+      "loss": 1.1152,
+      "step": 3
+    },
+    {
+      "epoch": 0.02099737532808399,
+      "grad_norm": 0.9984297481710986,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.0953,
+      "step": 4
+    },
+    {
+      "epoch": 0.026246719160104987,
+      "grad_norm": 0.8302026280344834,
+      "learning_rate": 5e-06,
+      "loss": 1.0617,
+      "step": 5
+    },
+    {
+      "epoch": 0.031496062992125984,
+      "grad_norm": 0.8911823807745126,
+      "learning_rate": 6e-06,
+      "loss": 1.1297,
+      "step": 6
+    },
+    {
+      "epoch": 0.03674540682414698,
+      "grad_norm": 0.686211615667355,
+      "learning_rate": 7e-06,
+      "loss": 1.0705,
+      "step": 7
+    },
+    {
+      "epoch": 0.04199475065616798,
+      "grad_norm": 0.9091855799181295,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.065,
+      "step": 8
+    },
+    {
+      "epoch": 0.047244094488188976,
+      "grad_norm": 0.8934722980371054,
+      "learning_rate": 9e-06,
+      "loss": 1.0767,
+      "step": 9
+    },
+    {
+      "epoch": 0.05249343832020997,
+      "grad_norm": 0.8688110393935611,
+      "learning_rate": 1e-05,
+      "loss": 1.0303,
+      "step": 10
+    },
+    {
+      "epoch": 0.05774278215223097,
+      "grad_norm": 0.9920393807379069,
+      "learning_rate": 1.1000000000000001e-05,
+      "loss": 1.0855,
+      "step": 11
+    },
+    {
+      "epoch": 0.06299212598425197,
+      "grad_norm": 0.9220245541797021,
+      "learning_rate": 1.2e-05,
+      "loss": 1.0531,
+      "step": 12
+    },
+    {
+      "epoch": 0.06824146981627296,
+      "grad_norm": 0.736886642754733,
+      "learning_rate": 1.3000000000000001e-05,
+      "loss": 1.0456,
+      "step": 13
+    },
+    {
+      "epoch": 0.07349081364829396,
+      "grad_norm": 0.771339891024354,
+      "learning_rate": 1.4e-05,
+      "loss": 1.0671,
+      "step": 14
+    },
+    {
+      "epoch": 0.07874015748031496,
+      "grad_norm": 0.7161080553611359,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.0521,
+      "step": 15
+    },
+    {
+      "epoch": 0.08398950131233596,
+      "grad_norm": 0.6788342613059561,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 1.0674,
+      "step": 16
+    },
+    {
+      "epoch": 0.08923884514435695,
+      "grad_norm": 0.7102848455414168,
+      "learning_rate": 1.7e-05,
+      "loss": 1.0459,
+      "step": 17
+    },
+    {
+      "epoch": 0.09448818897637795,
+      "grad_norm": 0.6425246555654909,
+      "learning_rate": 1.8e-05,
+      "loss": 1.0093,
+      "step": 18
+    },
+    {
+      "epoch": 0.09973753280839895,
+      "grad_norm": 2.099748819540086,
+      "learning_rate": 1.9e-05,
+      "loss": 1.0301,
+      "step": 19
+    },
+    {
+      "epoch": 0.10498687664041995,
+      "grad_norm": 0.6691987921672391,
+      "learning_rate": 2e-05,
+      "loss": 1.0199,
+      "step": 20
+    },
+    {
+      "epoch": 0.11023622047244094,
+      "grad_norm": 0.5883655485426926,
+      "learning_rate": 2.1000000000000002e-05,
+      "loss": 1.0085,
+      "step": 21
+    },
+    {
+      "epoch": 0.11548556430446194,
+      "grad_norm": 0.5443706631485103,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 1.0432,
+      "step": 22
+    },
+    {
+      "epoch": 0.12073490813648294,
+      "grad_norm": 0.593023936793411,
+      "learning_rate": 2.3e-05,
+      "loss": 1.0196,
+      "step": 23
+    },
+    {
+      "epoch": 0.12598425196850394,
+      "grad_norm": 0.5618656915734137,
+      "learning_rate": 2.4e-05,
+      "loss": 1.0386,
+      "step": 24
+    },
+    {
+      "epoch": 0.13123359580052493,
+      "grad_norm": 0.46871710098096486,
+      "learning_rate": 2.5e-05,
+      "loss": 0.9611,
+      "step": 25
+    },
+    {
+      "epoch": 0.13648293963254593,
+      "grad_norm": 0.5700902276763852,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 1.0045,
+      "step": 26
+    },
+    {
+      "epoch": 0.14173228346456693,
+      "grad_norm": 0.603692765386866,
+      "learning_rate": 2.7000000000000002e-05,
+      "loss": 1.019,
+      "step": 27
+    },
+    {
+      "epoch": 0.14698162729658792,
+      "grad_norm": 0.48456720859923497,
+      "learning_rate": 2.8e-05,
+      "loss": 0.9892,
+      "step": 28
+    },
+    {
+      "epoch": 0.15223097112860892,
+      "grad_norm": 0.45729475661677665,
+      "learning_rate": 2.9e-05,
+      "loss": 0.9645,
+      "step": 29
+    },
+    {
+      "epoch": 0.15748031496062992,
+      "grad_norm": 0.5439846777665153,
+      "learning_rate": 3.0000000000000004e-05,
+      "loss": 0.9497,
+      "step": 30
+    },
+    {
+      "epoch": 0.16272965879265092,
+      "grad_norm": 0.4965459941185334,
+      "learning_rate": 3.1e-05,
+      "loss": 0.9882,
+      "step": 31
+    },
+    {
+      "epoch": 0.1679790026246719,
+      "grad_norm": 0.4656328962534996,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 1.0057,
+      "step": 32
+    },
+    {
+      "epoch": 0.1732283464566929,
+      "grad_norm": 0.5241601609773927,
+      "learning_rate": 3.3e-05,
+      "loss": 1.0033,
+      "step": 33
+    },
+    {
+      "epoch": 0.1784776902887139,
+      "grad_norm": 0.5062226992393802,
+      "learning_rate": 3.4e-05,
+      "loss": 1.0166,
+      "step": 34
+    },
+    {
+      "epoch": 0.1837270341207349,
+      "grad_norm": 0.43771829747985674,
+      "learning_rate": 3.5000000000000004e-05,
+      "loss": 1.0102,
+      "step": 35
+    },
+    {
+      "epoch": 0.1889763779527559,
+      "grad_norm": 0.48092156639697076,
+      "learning_rate": 3.6e-05,
+      "loss": 1.018,
+      "step": 36
+    },
+    {
+      "epoch": 0.1942257217847769,
+      "grad_norm": 0.48115559949514536,
+      "learning_rate": 3.7000000000000005e-05,
+      "loss": 1.0079,
+      "step": 37
+    },
+    {
+      "epoch": 0.1994750656167979,
+      "grad_norm": 0.4777546937622387,
+      "learning_rate": 3.8e-05,
+      "loss": 1.0085,
+      "step": 38
+    },
+    {
+      "epoch": 0.2047244094488189,
+      "grad_norm": 0.44755392669080185,
+      "learning_rate": 3.9e-05,
+      "loss": 0.9825,
+      "step": 39
+    },
+    {
+      "epoch": 0.2099737532808399,
+      "grad_norm": 0.44510881962201315,
+      "learning_rate": 4e-05,
+      "loss": 0.9848,
+      "step": 40
+    },
+    {
+      "epoch": 0.2152230971128609,
+      "grad_norm": 0.4746290969046573,
+      "learning_rate": 3.999914623406736e-05,
+      "loss": 0.9888,
+      "step": 41
+    },
+    {
+      "epoch": 0.2204724409448819,
+      "grad_norm": 0.5953130701884418,
+      "learning_rate": 3.9996585009161056e-05,
+      "loss": 0.9882,
+      "step": 42
+    },
+    {
+      "epoch": 0.22572178477690288,
+      "grad_norm": 0.4251472611705547,
+      "learning_rate": 3.999231654394975e-05,
+      "loss": 0.9958,
+      "step": 43
+    },
+    {
+      "epoch": 0.23097112860892388,
+      "grad_norm": 0.44690799367073597,
+      "learning_rate": 3.9986341202860467e-05,
+      "loss": 0.9543,
+      "step": 44
+    },
+    {
+      "epoch": 0.23622047244094488,
+      "grad_norm": 0.5976579403936895,
+      "learning_rate": 3.9978659496047456e-05,
+      "loss": 0.9762,
+      "step": 45
+    },
+    {
+      "epoch": 0.24146981627296588,
+      "grad_norm": 0.3962092871428472,
+      "learning_rate": 3.9969272079348685e-05,
+      "loss": 0.9605,
+      "step": 46
+    },
+    {
+      "epoch": 0.24671916010498687,
+      "grad_norm": 0.43362883575028716,
+      "learning_rate": 3.995817975422981e-05,
+      "loss": 0.9456,
+      "step": 47
+    },
+    {
+      "epoch": 0.25196850393700787,
+      "grad_norm": 0.4139776793240363,
+      "learning_rate": 3.994538346771576e-05,
+      "loss": 0.9165,
+      "step": 48
+    },
+    {
+      "epoch": 0.2572178477690289,
+      "grad_norm": 0.3940723609427906,
+      "learning_rate": 3.9930884312309894e-05,
+      "loss": 0.9071,
+      "step": 49
+    },
+    {
+      "epoch": 0.26246719160104987,
+      "grad_norm": 0.4016006422322008,
+      "learning_rate": 3.991468352590069e-05,
+      "loss": 0.9668,
+      "step": 50
+    },
+    {
+      "epoch": 0.2677165354330709,
+      "grad_norm": 0.9528446542157881,
+      "learning_rate": 3.989678249165612e-05,
+      "loss": 1.0431,
+      "step": 51
+    },
+    {
+      "epoch": 0.27296587926509186,
+      "grad_norm": 0.41600529189619084,
+      "learning_rate": 3.987718273790548e-05,
+      "loss": 0.9464,
+      "step": 52
+    },
+    {
+      "epoch": 0.2782152230971129,
+      "grad_norm": 1.1382476752327089,
+      "learning_rate": 3.9855885938008986e-05,
+      "loss": 1.0186,
+      "step": 53
+    },
+    {
+      "epoch": 0.28346456692913385,
+      "grad_norm": 0.44849148754190465,
+      "learning_rate": 3.983289391021486e-05,
+      "loss": 0.9981,
+      "step": 54
+    },
+    {
+      "epoch": 0.2887139107611549,
+      "grad_norm": 0.4296819710357216,
+      "learning_rate": 3.9808208617504106e-05,
+      "loss": 0.9124,
+      "step": 55
+    },
+    {
+      "epoch": 0.29396325459317585,
+      "grad_norm": 1.4708100276334197,
+      "learning_rate": 3.9781832167422926e-05,
+      "loss": 1.0627,
+      "step": 56
+    },
+    {
+      "epoch": 0.2992125984251969,
+      "grad_norm": 0.436502847615945,
+      "learning_rate": 3.9753766811902756e-05,
+      "loss": 0.9399,
+      "step": 57
+    },
+    {
+      "epoch": 0.30446194225721784,
+      "grad_norm": 0.41131082586189677,
+      "learning_rate": 3.972401494706805e-05,
+      "loss": 0.9381,
+      "step": 58
+    },
+    {
+      "epoch": 0.30971128608923887,
+      "grad_norm": 0.42792569998778285,
+      "learning_rate": 3.969257911303167e-05,
+      "loss": 0.9426,
+      "step": 59
+    },
+    {
+      "epoch": 0.31496062992125984,
+      "grad_norm": 1.0484985550985957,
+      "learning_rate": 3.965946199367804e-05,
+      "loss": 1.0745,
+      "step": 60
+    },
+    {
+      "epoch": 0.32020997375328086,
+      "grad_norm": 0.45563925287513607,
+      "learning_rate": 3.962466641643398e-05,
+      "loss": 1.0085,
+      "step": 61
+    },
+    {
+      "epoch": 0.32545931758530183,
+      "grad_norm": 0.4216131864169055,
+      "learning_rate": 3.958819535202732e-05,
+      "loss": 0.9533,
+      "step": 62
+    },
+    {
+      "epoch": 0.33070866141732286,
+      "grad_norm": 0.47284588975540814,
+      "learning_rate": 3.9550051914233314e-05,
+      "loss": 0.9727,
+      "step": 63
+    },
+    {
+      "epoch": 0.3359580052493438,
+      "grad_norm": 0.4112493584955737,
+      "learning_rate": 3.951023935960874e-05,
+      "loss": 0.9408,
+      "step": 64
+    },
+    {
+      "epoch": 0.34120734908136485,
+      "grad_norm": 0.44123500755805545,
+      "learning_rate": 3.9468761087213864e-05,
+      "loss": 0.9547,
+      "step": 65
+    },
+    {
+      "epoch": 0.3464566929133858,
+      "grad_norm": 0.4160767709488051,
+      "learning_rate": 3.942562063832228e-05,
+      "loss": 0.9862,
+      "step": 66
+    },
+    {
+      "epoch": 0.35170603674540685,
+      "grad_norm": 0.40282812591350464,
+      "learning_rate": 3.9380821696118556e-05,
+      "loss": 0.9301,
+      "step": 67
+    },
+    {
+      "epoch": 0.3569553805774278,
+      "grad_norm": 0.42252313457664165,
+      "learning_rate": 3.933436808538375e-05,
+      "loss": 0.9751,
+      "step": 68
+    },
+    {
+      "epoch": 0.36220472440944884,
+      "grad_norm": 0.4084367556454159,
+      "learning_rate": 3.92862637721689e-05,
+      "loss": 0.9838,
+      "step": 69
+    },
+    {
+      "epoch": 0.3674540682414698,
+      "grad_norm": 0.39446053200993564,
+      "learning_rate": 3.923651286345638e-05,
+      "loss": 0.9237,
+      "step": 70
+    },
+    {
+      "epoch": 0.37270341207349084,
+      "grad_norm": 0.43051114259650114,
+      "learning_rate": 3.9185119606809305e-05,
+      "loss": 0.9543,
+      "step": 71
+    },
+    {
+      "epoch": 0.3779527559055118,
+      "grad_norm": 0.41527447901851827,
+      "learning_rate": 3.913208839000882e-05,
+      "loss": 0.9688,
+      "step": 72
+    },
+    {
+      "epoch": 0.38320209973753283,
+      "grad_norm": 0.4033220715509175,
+      "learning_rate": 3.907742374067956e-05,
+      "loss": 0.9401,
+      "step": 73
+    },
+    {
+      "epoch": 0.3884514435695538,
+      "grad_norm": 0.4039636146150166,
+      "learning_rate": 3.9021130325903076e-05,
+      "loss": 0.9621,
+      "step": 74
+    },
+    {
+      "epoch": 0.3937007874015748,
+      "grad_norm": 0.3896809489063709,
+      "learning_rate": 3.896321295181932e-05,
+      "loss": 0.986,
+      "step": 75
+    },
+    {
+      "epoch": 0.3989501312335958,
+      "grad_norm": 0.7547382513819603,
+      "learning_rate": 3.89036765632164e-05,
+      "loss": 1.0528,
+      "step": 76
+    },
+    {
+      "epoch": 0.4041994750656168,
+      "grad_norm": 0.42422582617937166,
+      "learning_rate": 3.8842526243108326e-05,
+      "loss": 0.9541,
+      "step": 77
+    },
+    {
+      "epoch": 0.4094488188976378,
+      "grad_norm": 0.41581388939730257,
+      "learning_rate": 3.877976721230114e-05,
+      "loss": 0.9711,
+      "step": 78
+    },
+    {
+      "epoch": 0.4146981627296588,
+      "grad_norm": 0.4326138308224312,
+      "learning_rate": 3.8715404828947055e-05,
+      "loss": 0.9261,
+      "step": 79
+    },
+    {
+      "epoch": 0.4199475065616798,
+      "grad_norm": 0.38852695749391314,
+      "learning_rate": 3.864944458808712e-05,
+      "loss": 0.9648,
+      "step": 80
+    },
+    {
+      "epoch": 0.4251968503937008,
+      "grad_norm": 0.3897195092049238,
+      "learning_rate": 3.8581892121181984e-05,
+      "loss": 0.9397,
+      "step": 81
+    },
+    {
+      "epoch": 0.4304461942257218,
+      "grad_norm": 0.43934794613481915,
+      "learning_rate": 3.851275319563113e-05,
+      "loss": 0.9905,
+      "step": 82
+    },
+    {
+      "epoch": 0.4356955380577428,
+      "grad_norm": 0.5323662587576004,
+      "learning_rate": 3.844203371428049e-05,
+      "loss": 0.9896,
+      "step": 83
+    },
+    {
+      "epoch": 0.4409448818897638,
+      "grad_norm": 0.38441956539336747,
+      "learning_rate": 3.836973971491847e-05,
+      "loss": 0.9385,
+      "step": 84
+    },
+    {
+      "epoch": 0.4461942257217848,
+      "grad_norm": 0.38662975914153885,
+      "learning_rate": 3.8295877369760426e-05,
+      "loss": 0.9586,
+      "step": 85
+    },
+    {
+      "epoch": 0.45144356955380577,
+      "grad_norm": 0.41009140101075614,
+      "learning_rate": 3.822045298492177e-05,
+      "loss": 0.9667,
+      "step": 86
+    },
+    {
+      "epoch": 0.4566929133858268,
+      "grad_norm": 0.4258642992742759,
+      "learning_rate": 3.814347299987953e-05,
+      "loss": 0.954,
+      "step": 87
+    },
+    {
+      "epoch": 0.46194225721784776,
+      "grad_norm": 0.40527142541860056,
+      "learning_rate": 3.806494398692258e-05,
+      "loss": 0.9351,
+      "step": 88
+    },
+    {
+      "epoch": 0.4671916010498688,
+      "grad_norm": 0.3743850574341336,
+      "learning_rate": 3.7984872650590516e-05,
+      "loss": 0.9498,
+      "step": 89
+    },
+    {
+      "epoch": 0.47244094488188976,
+      "grad_norm": 0.4151867667600151,
+      "learning_rate": 3.790326582710125e-05,
+      "loss": 0.9466,
+      "step": 90
+    },
+    {
+      "epoch": 0.4776902887139108,
+      "grad_norm": 0.4448011376311795,
+      "learning_rate": 3.782013048376736e-05,
+      "loss": 1.0266,
+      "step": 91
+    },
+    {
+      "epoch": 0.48293963254593175,
+      "grad_norm": 0.38192124855359877,
+      "learning_rate": 3.773547371840124e-05,
+      "loss": 0.978,
+      "step": 92
+    },
+    {
+      "epoch": 0.4881889763779528,
+      "grad_norm": 0.4235778210861527,
+      "learning_rate": 3.764930275870912e-05,
+      "loss": 0.9827,
+      "step": 93
+    },
+    {
+      "epoch": 0.49343832020997375,
+      "grad_norm": 0.4051195260626496,
+      "learning_rate": 3.756162496167396e-05,
+      "loss": 0.963,
+      "step": 94
+    },
+    {
+      "epoch": 0.49868766404199477,
+      "grad_norm": 0.40700055373961197,
+      "learning_rate": 3.7472447812927395e-05,
+      "loss": 0.9437,
+      "step": 95
+    },
+    {
+      "epoch": 0.5039370078740157,
+      "grad_norm": 0.38712614108502513,
+      "learning_rate": 3.738177892611057e-05,
+      "loss": 0.955,
+      "step": 96
+    },
+    {
+      "epoch": 0.5091863517060368,
+      "grad_norm": 0.4099596350735423,
+      "learning_rate": 3.728962604222416e-05,
+      "loss": 0.9741,
+      "step": 97
+    },
+    {
+      "epoch": 0.5144356955380578,
+      "grad_norm": 0.40040635119594403,
+      "learning_rate": 3.719599702896745e-05,
+      "loss": 0.9528,
+      "step": 98
+    },
+    {
+      "epoch": 0.5196850393700787,
+      "grad_norm": 0.4136053200425271,
+      "learning_rate": 3.710089988006662e-05,
+      "loss": 0.9466,
+      "step": 99
+    },
+    {
+      "epoch": 0.5249343832020997,
+      "grad_norm": 0.41412239719227456,
+      "learning_rate": 3.700434271459229e-05,
+      "loss": 0.9242,
+      "step": 100
+    },
+    {
+      "epoch": 0.5301837270341208,
+      "grad_norm": 0.4309979528684408,
+      "learning_rate": 3.690633377626628e-05,
+      "loss": 0.9861,
+      "step": 101
+    },
+    {
+      "epoch": 0.5354330708661418,
+      "grad_norm": 0.4064293156979199,
+      "learning_rate": 3.680688143275786e-05,
+      "loss": 0.931,
+      "step": 102
+    },
+    {
+      "epoch": 0.5406824146981627,
+      "grad_norm": 0.4463450853160405,
+      "learning_rate": 3.670599417496931e-05,
+      "loss": 0.9084,
+      "step": 103
+    },
+    {
+      "epoch": 0.5459317585301837,
+      "grad_norm": 0.4542877579158036,
+      "learning_rate": 3.6603680616311013e-05,
+      "loss": 0.9561,
+      "step": 104
+    },
+    {
+      "epoch": 0.5511811023622047,
+      "grad_norm": 0.4606576229715047,
+      "learning_rate": 3.6499949491966046e-05,
+      "loss": 0.9424,
+      "step": 105
+    },
+    {
+      "epoch": 0.5564304461942258,
+      "grad_norm": 1.6662857295077933,
+      "learning_rate": 3.639480965814443e-05,
+      "loss": 1.0371,
+      "step": 106
+    },
+    {
+      "epoch": 0.5616797900262467,
+      "grad_norm": 0.42684188670392853,
+      "learning_rate": 3.628827009132697e-05,
+      "loss": 0.9635,
+      "step": 107
+    },
+    {
+      "epoch": 0.5669291338582677,
+      "grad_norm": 1.2208350090054685,
+      "learning_rate": 3.6180339887498953e-05,
+      "loss": 0.9917,
+      "step": 108
+    },
+    {
+      "epoch": 0.5721784776902887,
+      "grad_norm": 0.4294502318914682,
+      "learning_rate": 3.6071028261373474e-05,
+      "loss": 0.9446,
+      "step": 109
+    },
+    {
+      "epoch": 0.5774278215223098,
+      "grad_norm": 0.3937562720593612,
+      "learning_rate": 3.5960344545604796e-05,
+      "loss": 0.9278,
+      "step": 110
+    },
+    {
+      "epoch": 0.5826771653543307,
+      "grad_norm": 1.4854854417438403,
+      "learning_rate": 3.584829818999148e-05,
+      "loss": 1.0161,
+      "step": 111
+    },
+    {
+      "epoch": 0.5879265091863517,
+      "grad_norm": 0.4240627994414154,
+      "learning_rate": 3.573489876066967e-05,
+      "loss": 0.9483,
+      "step": 112
+    },
+    {
+      "epoch": 0.5931758530183727,
+      "grad_norm": 0.3995864923040328,
+      "learning_rate": 3.5620155939296314e-05,
+      "loss": 0.9426,
+      "step": 113
+    },
+    {
+      "epoch": 0.5984251968503937,
+      "grad_norm": 0.4085167442197417,
+      "learning_rate": 3.55040795222226e-05,
+      "loss": 0.9189,
+      "step": 114
+    },
+    {
+      "epoch": 0.6036745406824147,
+      "grad_norm": 0.411605976954782,
+      "learning_rate": 3.538667941965758e-05,
+      "loss": 0.9406,
+      "step": 115
+    },
+    {
+      "epoch": 0.6089238845144357,
+      "grad_norm": 0.4510885035850897,
+      "learning_rate": 3.526796565482206e-05,
+      "loss": 0.9609,
+      "step": 116
+    },
+    {
+      "epoch": 0.6141732283464567,
+      "grad_norm": 0.39711542861711363,
+      "learning_rate": 3.514794836309286e-05,
+      "loss": 0.9353,
+      "step": 117
+    },
+    {
+      "epoch": 0.6194225721784777,
+      "grad_norm": 0.3860750426711258,
+      "learning_rate": 3.502663779113747e-05,
+      "loss": 0.9168,
+      "step": 118
+    },
+    {
+      "epoch": 0.6246719160104987,
+      "grad_norm": 0.4324143866853257,
+      "learning_rate": 3.490404429603925e-05,
+      "loss": 0.9412,
+      "step": 119
+    },
+    {
+      "epoch": 0.6299212598425197,
+      "grad_norm": 0.42486288700695524,
+      "learning_rate": 3.478017834441319e-05,
+      "loss": 0.9967,
+      "step": 120
+    },
+    {
+      "epoch": 0.6351706036745407,
+      "grad_norm": 0.42059534343716903,
+      "learning_rate": 3.4655050511512236e-05,
+      "loss": 0.9042,
+      "step": 121
+    },
+    {
+      "epoch": 0.6404199475065617,
+      "grad_norm": 0.375540386715667,
+      "learning_rate": 3.452867148032449e-05,
+      "loss": 0.9261,
+      "step": 122
+    },
+    {
+      "epoch": 0.6456692913385826,
+      "grad_norm": 0.38698966212541075,
+      "learning_rate": 3.44010520406611e-05,
+      "loss": 0.9252,
+      "step": 123
+    },
+    {
+      "epoch": 0.6509186351706037,
+      "grad_norm": 0.41709615104288367,
+      "learning_rate": 3.427220308823505e-05,
+      "loss": 0.9253,
+      "step": 124
+    },
+    {
+      "epoch": 0.6561679790026247,
+      "grad_norm": 0.4293707133542124,
+      "learning_rate": 3.4142135623730954e-05,
+      "loss": 0.9545,
+      "step": 125
+    },
+    {
+      "epoch": 0.6614173228346457,
+      "grad_norm": 0.40563024635145306,
+      "learning_rate": 3.401086075186582e-05,
+      "loss": 0.9424,
+      "step": 126
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.47226124502094396,
+      "learning_rate": 3.3878389680440995e-05,
+      "loss": 0.9408,
+      "step": 127
+    },
+    {
+      "epoch": 0.6719160104986877,
+      "grad_norm": 0.3921360030995963,
+      "learning_rate": 3.374473371938526e-05,
+      "loss": 0.9309,
+      "step": 128
+    },
+    {
+      "epoch": 0.6771653543307087,
+      "grad_norm": 0.4188603496902975,
+      "learning_rate": 3.3609904279789235e-05,
+      "loss": 0.9625,
+      "step": 129
+    },
+    {
+      "epoch": 0.6824146981627297,
+      "grad_norm": 0.40729320283126413,
+      "learning_rate": 3.347391287293115e-05,
+      "loss": 0.9222,
+      "step": 130
+    },
+    {
+      "epoch": 0.6876640419947506,
+      "grad_norm": 0.43355828675253894,
+      "learning_rate": 3.333677110929403e-05,
+      "loss": 0.9245,
+      "step": 131
+    },
+    {
+      "epoch": 0.6929133858267716,
+      "grad_norm": 0.40875412645403303,
+      "learning_rate": 3.319849069757446e-05,
+      "loss": 0.9416,
+      "step": 132
+    },
+    {
+      "epoch": 0.6981627296587927,
+      "grad_norm": 0.4184583102080097,
+      "learning_rate": 3.305908344368289e-05,
+      "loss": 0.9575,
+      "step": 133
+    },
+    {
+      "epoch": 0.7034120734908137,
+      "grad_norm": 0.37949729176161695,
+      "learning_rate": 3.291856124973575e-05,
+      "loss": 0.9283,
+      "step": 134
+    },
+    {
+      "epoch": 0.7086614173228346,
+      "grad_norm": 0.4359197990076154,
+      "learning_rate": 3.277693611303922e-05,
+      "loss": 0.9591,
+      "step": 135
+    },
+    {
+      "epoch": 0.7139107611548556,
+      "grad_norm": 0.4127988509227564,
+      "learning_rate": 3.263422012506502e-05,
+      "loss": 0.9507,
+      "step": 136
+    },
+    {
+      "epoch": 0.7191601049868767,
+      "grad_norm": 0.4119681718108907,
+      "learning_rate": 3.249042547041799e-05,
+      "loss": 0.9252,
+      "step": 137
+    },
+    {
+      "epoch": 0.7244094488188977,
+      "grad_norm": 0.4155554867266832,
+      "learning_rate": 3.234556442579586e-05,
+      "loss": 0.9263,
+      "step": 138
+    },
+    {
+      "epoch": 0.7296587926509186,
+      "grad_norm": 0.37277040517135684,
+      "learning_rate": 3.219964935894114e-05,
+      "loss": 0.9544,
+      "step": 139
+    },
+    {
+      "epoch": 0.7349081364829396,
+      "grad_norm": 0.41745861140292206,
+      "learning_rate": 3.205269272758513e-05,
+      "loss": 0.9213,
+      "step": 140
+    },
+    {
+      "epoch": 0.7401574803149606,
+      "grad_norm": 0.41788351218514774,
+      "learning_rate": 3.190470707838438e-05,
+      "loss": 0.9429,
+      "step": 141
+    },
+    {
+      "epoch": 0.7454068241469817,
+      "grad_norm": 0.3994620013935183,
+      "learning_rate": 3.1755705045849465e-05,
+      "loss": 0.9065,
+      "step": 142
+    },
+    {
+      "epoch": 0.7506561679790026,
+      "grad_norm": 0.4006844018528632,
+      "learning_rate": 3.160569935126632e-05,
+      "loss": 0.9064,
+      "step": 143
+    },
+    {
+      "epoch": 0.7559055118110236,
+      "grad_norm": 0.44223134289541643,
+      "learning_rate": 3.145470280161011e-05,
+      "loss": 0.9247,
+      "step": 144
+    },
+    {
+      "epoch": 0.7611548556430446,
+      "grad_norm": 0.41494829719611687,
+      "learning_rate": 3.130272828845184e-05,
+      "loss": 0.9126,
+      "step": 145
+    },
+    {
+      "epoch": 0.7664041994750657,
+      "grad_norm": 0.38947944768031434,
+      "learning_rate": 3.114978878685771e-05,
+      "loss": 0.8928,
+      "step": 146
+    },
+    {
+      "epoch": 0.7716535433070866,
+      "grad_norm": 0.3945361927140775,
+      "learning_rate": 3.0995897354281347e-05,
+      "loss": 0.889,
+      "step": 147
+    },
+    {
+      "epoch": 0.7769028871391076,
+      "grad_norm": 0.39978716157020916,
+      "learning_rate": 3.084106712944899e-05,
+      "loss": 0.9227,
+      "step": 148
+    },
+    {
+      "epoch": 0.7821522309711286,
+      "grad_norm": 0.39603745657551037,
+      "learning_rate": 3.068531133123777e-05,
+      "loss": 0.8969,
+      "step": 149
+    },
+    {
+      "epoch": 0.7874015748031497,
+      "grad_norm": 1.3504046960889928,
+      "learning_rate": 3.052864325754712e-05,
+      "loss": 1.0631,
+      "step": 150
+    },
+    {
+      "epoch": 0.7926509186351706,
+      "grad_norm": 0.4310383398496922,
+      "learning_rate": 3.0371076284163442e-05,
+      "loss": 0.9262,
+      "step": 151
+    },
+    {
+      "epoch": 0.7979002624671916,
+      "grad_norm": 0.41699772424137066,
+      "learning_rate": 3.021262386361814e-05,
+      "loss": 0.9352,
+      "step": 152
+    },
+    {
+      "epoch": 0.8031496062992126,
+      "grad_norm": 0.4056852584293386,
+      "learning_rate": 3.0053299524039077e-05,
+      "loss": 0.8957,
+      "step": 153
+    },
+    {
+      "epoch": 0.8083989501312336,
+      "grad_norm": 0.4308645558537417,
+      "learning_rate": 2.9893116867995583e-05,
+      "loss": 0.9137,
+      "step": 154
+    },
+    {
+      "epoch": 0.8136482939632546,
+      "grad_norm": 0.39136699559712107,
+      "learning_rate": 2.9732089571337126e-05,
+      "loss": 0.9392,
+      "step": 155
+    },
+    {
+      "epoch": 0.8188976377952756,
+      "grad_norm": 0.39692286867805615,
+      "learning_rate": 2.9570231382025732e-05,
+      "loss": 0.9319,
+      "step": 156
+    },
+    {
+      "epoch": 0.8241469816272966,
+      "grad_norm": 0.389760753952324,
+      "learning_rate": 2.9407556118962192e-05,
+      "loss": 0.9328,
+      "step": 157
+    },
+    {
+      "epoch": 0.8293963254593176,
+      "grad_norm": 0.40644738344754366,
+      "learning_rate": 2.924407767080627e-05,
+      "loss": 0.9511,
+      "step": 158
+    },
+    {
+      "epoch": 0.8346456692913385,
+      "grad_norm": 0.4235598803780184,
+      "learning_rate": 2.9079809994790937e-05,
+      "loss": 0.9443,
+      "step": 159
+    },
+    {
+      "epoch": 0.8398950131233596,
+      "grad_norm": 0.39469735698768543,
+      "learning_rate": 2.891476711553077e-05,
+      "loss": 0.9353,
+      "step": 160
+    },
+    {
+      "epoch": 0.8451443569553806,
+      "grad_norm": 0.4231486830962651,
+      "learning_rate": 2.8748963123824532e-05,
+      "loss": 0.9598,
+      "step": 161
+    },
+    {
+      "epoch": 0.8503937007874016,
+      "grad_norm": 0.4016499332546737,
+      "learning_rate": 2.858241217545218e-05,
+      "loss": 0.9182,
+      "step": 162
+    },
+    {
+      "epoch": 0.8556430446194225,
+      "grad_norm": 0.7416569697844047,
+      "learning_rate": 2.8415128489966308e-05,
+      "loss": 1.017,
+      "step": 163
+    },
+    {
+      "epoch": 0.8608923884514436,
+      "grad_norm": 0.4049886957012087,
+      "learning_rate": 2.8247126349478073e-05,
+      "loss": 0.9377,
+      "step": 164
+    },
+    {
+      "epoch": 0.8661417322834646,
+      "grad_norm": 0.4240641122781046,
+      "learning_rate": 2.80784200974379e-05,
+      "loss": 0.936,
+      "step": 165
+    },
+    {
+      "epoch": 0.8713910761154856,
+      "grad_norm": 0.4010428790320475,
+      "learning_rate": 2.790902413741085e-05,
+      "loss": 0.9076,
+      "step": 166
+    },
+    {
+      "epoch": 0.8766404199475065,
+      "grad_norm": 0.40515062001849617,
+      "learning_rate": 2.773895293184691e-05,
+      "loss": 0.9144,
+      "step": 167
+    },
+    {
+      "epoch": 0.8818897637795275,
+      "grad_norm": 0.4171752905975984,
+      "learning_rate": 2.756822100084621e-05,
+      "loss": 0.9302,
+      "step": 168
+    },
+    {
+      "epoch": 0.8871391076115486,
+      "grad_norm": 0.4018514009140958,
+      "learning_rate": 2.7396842920919384e-05,
+      "loss": 0.9208,
+      "step": 169
+    },
+    {
+      "epoch": 0.8923884514435696,
+      "grad_norm": 0.39277733117068253,
+      "learning_rate": 2.7224833323743064e-05,
+      "loss": 0.9116,
+      "step": 170
+    },
+    {
+      "epoch": 0.8976377952755905,
+      "grad_norm": 0.6692602521355003,
+      "learning_rate": 2.7052206894910653e-05,
+      "loss": 1.0122,
+      "step": 171
+    },
+    {
+      "epoch": 0.9028871391076115,
+      "grad_norm": 0.40843018046933677,
+      "learning_rate": 2.6878978372678567e-05,
+      "loss": 0.9014,
+      "step": 172
+    },
+    {
+      "epoch": 0.9081364829396326,
+      "grad_norm": 0.3862093081092539,
+      "learning_rate": 2.670516254670788e-05,
+      "loss": 0.9367,
+      "step": 173
+    },
+    {
+      "epoch": 0.9133858267716536,
+      "grad_norm": 0.39106222738031376,
+      "learning_rate": 2.6530774256801666e-05,
+      "loss": 0.9253,
+      "step": 174
+    },
+    {
+      "epoch": 0.9186351706036745,
+      "grad_norm": 0.409656789286683,
+      "learning_rate": 2.6355828391638036e-05,
+      "loss": 0.9259,
+      "step": 175
+    },
+    {
+      "epoch": 0.9238845144356955,
+      "grad_norm": 0.41048542482358136,
+      "learning_rate": 2.618033988749895e-05,
+      "loss": 0.9151,
+      "step": 176
+    },
+    {
+      "epoch": 0.9291338582677166,
+      "grad_norm": 0.39649048041899354,
+      "learning_rate": 2.6004323726995057e-05,
+      "loss": 0.9197,
+      "step": 177
+    },
+    {
+      "epoch": 0.9343832020997376,
+      "grad_norm": 0.4041163682720282,
+      "learning_rate": 2.5827794937786497e-05,
+      "loss": 0.9184,
+      "step": 178
+    },
+    {
+      "epoch": 0.9396325459317585,
+      "grad_norm": 0.40988870079100986,
+      "learning_rate": 2.5650768591299905e-05,
+      "loss": 0.9376,
+      "step": 179
+    },
+    {
+      "epoch": 0.9448818897637795,
+      "grad_norm": 0.3918596025187431,
+      "learning_rate": 2.5473259801441663e-05,
+      "loss": 0.9102,
+      "step": 180
+    },
+    {
+      "epoch": 0.9501312335958005,
+      "grad_norm": 0.39082215171197704,
+      "learning_rate": 2.5295283723307517e-05,
+      "loss": 0.9025,
+      "step": 181
+    },
+    {
+      "epoch": 0.9553805774278216,
+      "grad_norm": 0.38010414440929924,
+      "learning_rate": 2.5116855551888715e-05,
+      "loss": 0.9354,
+      "step": 182
+    },
+    {
+      "epoch": 0.9606299212598425,
+      "grad_norm": 0.4141554447250008,
+      "learning_rate": 2.4937990520774664e-05,
+      "loss": 0.8782,
+      "step": 183
+    },
+    {
+      "epoch": 0.9658792650918635,
+      "grad_norm": 0.38201600299774646,
+      "learning_rate": 2.4758703900852376e-05,
+      "loss": 0.9008,
+      "step": 184
+    },
+    {
+      "epoch": 0.9711286089238845,
+      "grad_norm": 0.42204019171609175,
+      "learning_rate": 2.4579010999002683e-05,
+      "loss": 0.8856,
+      "step": 185
+    },
+    {
+      "epoch": 0.9763779527559056,
+      "grad_norm": 0.4270135581368761,
+      "learning_rate": 2.4398927156793376e-05,
+      "loss": 0.9205,
+      "step": 186
+    },
+    {
+      "epoch": 0.9816272965879265,
+      "grad_norm": 0.4150364763728507,
+      "learning_rate": 2.42184677491694e-05,
+      "loss": 0.8947,
+      "step": 187
+    },
+    {
+      "epoch": 0.9868766404199475,
+      "grad_norm": 0.51571681852072,
+      "learning_rate": 2.4037648183140205e-05,
+      "loss": 0.9929,
+      "step": 188
+    },
+    {
+      "epoch": 0.9921259842519685,
+      "grad_norm": 0.38917079851953085,
+      "learning_rate": 2.385648389646434e-05,
+      "loss": 0.9121,
+      "step": 189
+    },
+    {
+      "epoch": 0.9973753280839895,
+      "grad_norm": 0.45530540311855816,
+      "learning_rate": 2.367499035633141e-05,
+      "loss": 0.9113,
+      "step": 190
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.45530540311855816,
+      "learning_rate": 2.3493183058041578e-05,
+      "loss": 0.9347,
+      "step": 191
+    },
+    {
+      "epoch": 1.005249343832021,
+      "grad_norm": 0.7905516160996418,
+      "learning_rate": 2.33110775236826e-05,
+      "loss": 0.6378,
+      "step": 192
+    },
+    {
+      "epoch": 1.010498687664042,
+      "grad_norm": 0.5781672455733289,
+      "learning_rate": 2.312868930080462e-05,
+      "loss": 0.639,
+      "step": 193
+    },
+    {
+      "epoch": 1.015748031496063,
+      "grad_norm": 1.0363184255291382,
+      "learning_rate": 2.2946033961092754e-05,
+      "loss": 0.6442,
+      "step": 194
+    },
+    {
+      "epoch": 1.020997375328084,
+      "grad_norm": 0.4833222962879287,
+      "learning_rate": 2.2763127099037646e-05,
+      "loss": 0.6441,
+      "step": 195
+    },
+    {
+      "epoch": 1.026246719160105,
+      "grad_norm": 0.5713369686471519,
+      "learning_rate": 2.257998433060407e-05,
+      "loss": 0.6667,
+      "step": 196
+    },
+    {
+      "epoch": 1.031496062992126,
+      "grad_norm": 0.4764665095808476,
+      "learning_rate": 2.2396621291897666e-05,
+      "loss": 0.6407,
+      "step": 197
+    },
+    {
+      "epoch": 1.036745406824147,
+      "grad_norm": 0.4588405259893301,
+      "learning_rate": 2.2213053637830016e-05,
+      "loss": 0.6146,
+      "step": 198
+    },
+    {
+      "epoch": 1.041994750656168,
+      "grad_norm": 0.42224522477353676,
+      "learning_rate": 2.2029297040782063e-05,
+      "loss": 0.6108,
+      "step": 199
+    },
+    {
+      "epoch": 1.047244094488189,
+      "grad_norm": 0.43495470036978195,
+      "learning_rate": 2.184536718926604e-05,
+      "loss": 0.6091,
+      "step": 200
+    },
+    {
+      "epoch": 1.05249343832021,
+      "grad_norm": 0.3908883372852395,
+      "learning_rate": 2.166127978658608e-05,
+      "loss": 0.618,
+      "step": 201
+    },
+    {
+      "epoch": 1.057742782152231,
+      "grad_norm": 0.41668658316901935,
+      "learning_rate": 2.147705054949748e-05,
+      "loss": 0.6345,
+      "step": 202
+    },
+    {
+      "epoch": 1.0629921259842519,
+      "grad_norm": 0.38162610574666733,
+      "learning_rate": 2.1292695206864887e-05,
+      "loss": 0.6077,
+      "step": 203
+    },
+    {
+      "epoch": 1.068241469816273,
+      "grad_norm": 0.359243479327879,
+      "learning_rate": 2.11082294983194e-05,
+      "loss": 0.6008,
+      "step": 204
+    },
+    {
+      "epoch": 1.073490813648294,
+      "grad_norm": 0.338388070214759,
+      "learning_rate": 2.0923669172914796e-05,
+      "loss": 0.6198,
+      "step": 205
+    },
+    {
+      "epoch": 1.078740157480315,
+      "grad_norm": 0.3736442857329487,
+      "learning_rate": 2.0739029987782903e-05,
+      "loss": 0.6038,
+      "step": 206
+    },
+    {
+      "epoch": 1.083989501312336,
+      "grad_norm": 0.38343041994799376,
+      "learning_rate": 2.055432770678833e-05,
+      "loss": 0.6283,
+      "step": 207
+    },
+    {
+      "epoch": 1.0892388451443569,
+      "grad_norm": 0.3839775235441468,
+      "learning_rate": 2.03695780991826e-05,
+      "loss": 0.621,
+      "step": 208
+    },
+    {
+      "epoch": 1.094488188976378,
+      "grad_norm": 0.4026675034831997,
+      "learning_rate": 2.018479693825782e-05,
+      "loss": 0.5967,
+      "step": 209
+    },
+    {
+      "epoch": 1.099737532808399,
+      "grad_norm": 0.354604482258031,
+      "learning_rate": 2e-05,
+      "loss": 0.5747,
+      "step": 210
+    },
+    {
+      "epoch": 1.10498687664042,
+      "grad_norm": 0.37448103017465234,
+      "learning_rate": 1.9815203061742188e-05,
+      "loss": 0.6207,
+      "step": 211
+    },
+    {
+      "epoch": 1.110236220472441,
+      "grad_norm": 0.3731173545168172,
+      "learning_rate": 1.9630421900817407e-05,
+      "loss": 0.6658,
+      "step": 212
+    },
+    {
+      "epoch": 1.1154855643044619,
+      "grad_norm": 0.4440821296075456,
+      "learning_rate": 1.9445672293211675e-05,
+      "loss": 0.6147,
+      "step": 213
+    },
+    {
+      "epoch": 1.120734908136483,
+      "grad_norm": 0.37062388694479104,
+      "learning_rate": 1.9260970012217107e-05,
+      "loss": 0.6235,
+      "step": 214
+    },
+    {
+      "epoch": 1.125984251968504,
+      "grad_norm": 0.352789345485834,
+      "learning_rate": 1.9076330827085214e-05,
+      "loss": 0.6186,
+      "step": 215
+    },
+    {
+      "epoch": 1.1312335958005248,
+      "grad_norm": 0.3510249370323435,
+      "learning_rate": 1.8891770501680602e-05,
+      "loss": 0.5919,
+      "step": 216
+    },
+    {
+      "epoch": 1.136482939632546,
+      "grad_norm": 0.36641452500356003,
+      "learning_rate": 1.8707304793135117e-05,
+      "loss": 0.6325,
+      "step": 217
+    },
+    {
+      "epoch": 1.141732283464567,
+      "grad_norm": 0.46538331418150264,
+      "learning_rate": 1.8522949450502522e-05,
+      "loss": 0.6314,
+      "step": 218
+    },
+    {
+      "epoch": 1.1469816272965878,
+      "grad_norm": 0.35158983297625845,
+      "learning_rate": 1.8338720213413924e-05,
+      "loss": 0.6171,
+      "step": 219
+    },
+    {
+      "epoch": 1.152230971128609,
+      "grad_norm": 0.34039327733619856,
+      "learning_rate": 1.815463281073396e-05,
+      "loss": 0.5929,
+      "step": 220
+    },
+    {
+      "epoch": 1.1574803149606299,
+      "grad_norm": 0.3583390822597362,
+      "learning_rate": 1.7970702959217944e-05,
+      "loss": 0.5666,
+      "step": 221
+    },
+    {
+      "epoch": 1.162729658792651,
+      "grad_norm": 0.3328969811354483,
+      "learning_rate": 1.7786946362169987e-05,
+      "loss": 0.6165,
+      "step": 222
+    },
+    {
+      "epoch": 1.167979002624672,
+      "grad_norm": 0.35808562196234894,
+      "learning_rate": 1.760337870810234e-05,
+      "loss": 0.5907,
+      "step": 223
+    },
+    {
+      "epoch": 1.1732283464566928,
+      "grad_norm": 0.3330883120593054,
+      "learning_rate": 1.742001566939594e-05,
+      "loss": 0.6139,
+      "step": 224
+    },
+    {
+      "epoch": 1.178477690288714,
+      "grad_norm": 0.3477344689492317,
+      "learning_rate": 1.7236872900962364e-05,
+      "loss": 0.5772,
+      "step": 225
+    },
+    {
+      "epoch": 1.1837270341207349,
+      "grad_norm": 0.3636049536281939,
+      "learning_rate": 1.705396603890725e-05,
+      "loss": 0.6101,
+      "step": 226
+    },
+    {
+      "epoch": 1.188976377952756,
+      "grad_norm": 0.5140813534017206,
+      "learning_rate": 1.687131069919538e-05,
+      "loss": 0.6455,
+      "step": 227
+    },
+    {
+      "epoch": 1.194225721784777,
+      "grad_norm": 0.3445915498918541,
+      "learning_rate": 1.66889224763174e-05,
+      "loss": 0.6253,
+      "step": 228
+    },
+    {
+      "epoch": 1.1994750656167978,
+      "grad_norm": 0.3817825400306628,
+      "learning_rate": 1.6506816941958425e-05,
+      "loss": 0.6264,
+      "step": 229
+    },
+    {
+      "epoch": 1.204724409448819,
+      "grad_norm": 0.3871021865440479,
+      "learning_rate": 1.6325009643668592e-05,
+      "loss": 0.5875,
+      "step": 230
+    },
+    {
+      "epoch": 1.20997375328084,
+      "grad_norm": 0.4003951939816748,
+      "learning_rate": 1.6143516103535666e-05,
+      "loss": 0.6068,
+      "step": 231
+    },
+    {
+      "epoch": 1.2152230971128608,
+      "grad_norm": 0.377400380096938,
+      "learning_rate": 1.59623518168598e-05,
+      "loss": 0.6216,
+      "step": 232
+    },
+    {
+      "epoch": 1.220472440944882,
+      "grad_norm": 0.3523802116183249,
+      "learning_rate": 1.578153225083061e-05,
+      "loss": 0.5773,
+      "step": 233
+    },
+    {
+      "epoch": 1.2257217847769029,
+      "grad_norm": 0.3774158319806266,
+      "learning_rate": 1.5601072843206634e-05,
+      "loss": 0.6485,
+      "step": 234
+    },
+    {
+      "epoch": 1.2309711286089238,
+      "grad_norm": 0.35618682756111997,
+      "learning_rate": 1.5420989000997324e-05,
+      "loss": 0.57,
+      "step": 235
+    },
+    {
+      "epoch": 1.236220472440945,
+      "grad_norm": 0.329939498867074,
+      "learning_rate": 1.524129609914763e-05,
+      "loss": 0.5922,
+      "step": 236
+    },
+    {
+      "epoch": 1.2414698162729658,
+      "grad_norm": 0.3694137451232253,
+      "learning_rate": 1.5062009479225336e-05,
+      "loss": 0.614,
+      "step": 237
+    },
+    {
+      "epoch": 1.246719160104987,
+      "grad_norm": 0.33955880982317405,
+      "learning_rate": 1.4883144448111288e-05,
+      "loss": 0.5734,
+      "step": 238
+    },
+    {
+      "epoch": 1.2519685039370079,
+      "grad_norm": 0.34083473110033147,
+      "learning_rate": 1.4704716276692483e-05,
+      "loss": 0.5838,
+      "step": 239
+    },
+    {
+      "epoch": 1.257217847769029,
+      "grad_norm": 0.33904875318136807,
+      "learning_rate": 1.4526740198558345e-05,
+      "loss": 0.5721,
+      "step": 240
+    },
+    {
+      "epoch": 1.26246719160105,
+      "grad_norm": 0.3452861169118219,
+      "learning_rate": 1.43492314087001e-05,
+      "loss": 0.5978,
+      "step": 241
+    },
+    {
+      "epoch": 1.2677165354330708,
+      "grad_norm": 0.36203138569672677,
+      "learning_rate": 1.417220506221351e-05,
+      "loss": 0.6061,
+      "step": 242
+    },
+    {
+      "epoch": 1.272965879265092,
+      "grad_norm": 0.3423141141716414,
+      "learning_rate": 1.3995676273004948e-05,
+      "loss": 0.617,
+      "step": 243
+    },
+    {
+      "epoch": 1.2782152230971129,
+      "grad_norm": 0.35751648695826527,
+      "learning_rate": 1.3819660112501054e-05,
+      "loss": 0.6393,
+      "step": 244
+    },
+    {
+      "epoch": 1.2834645669291338,
+      "grad_norm": 0.3703820503916198,
+      "learning_rate": 1.364417160836197e-05,
+      "loss": 0.6767,
+      "step": 245
+    },
+    {
+      "epoch": 1.288713910761155,
+      "grad_norm": 0.48807956046543455,
+      "learning_rate": 1.3469225743198337e-05,
+      "loss": 0.6156,
+      "step": 246
+    },
+    {
+      "epoch": 1.2939632545931758,
+      "grad_norm": 0.3687500146311602,
+      "learning_rate": 1.329483745329213e-05,
+      "loss": 0.6587,
+      "step": 247
+    },
+    {
+      "epoch": 1.2992125984251968,
+      "grad_norm": 0.5835964568778699,
+      "learning_rate": 1.3121021627321438e-05,
+      "loss": 0.5912,
+      "step": 248
+    },
+    {
+      "epoch": 1.304461942257218,
+      "grad_norm": 0.35312287301928247,
+      "learning_rate": 1.2947793105089347e-05,
+      "loss": 0.622,
+      "step": 249
+    },
+    {
+      "epoch": 1.3097112860892388,
+      "grad_norm": 0.3442037878519564,
+      "learning_rate": 1.2775166676256942e-05,
+      "loss": 0.5905,
+      "step": 250
+    },
+    {
+      "epoch": 1.3149606299212597,
+      "grad_norm": 0.3407336842366384,
+      "learning_rate": 1.260315707908062e-05,
+      "loss": 0.6023,
+      "step": 251
+    },
+    {
+      "epoch": 1.3202099737532809,
+      "grad_norm": 0.3532792670283708,
+      "learning_rate": 1.2431778999153796e-05,
+      "loss": 0.5994,
+      "step": 252
+    },
+    {
+      "epoch": 1.3254593175853018,
+      "grad_norm": 0.33944334283473854,
+      "learning_rate": 1.2261047068153098e-05,
+      "loss": 0.6136,
+      "step": 253
+    },
+    {
+      "epoch": 1.330708661417323,
+      "grad_norm": 0.32837150028469475,
+      "learning_rate": 1.2090975862589151e-05,
+      "loss": 0.6655,
+      "step": 254
+    },
+    {
+      "epoch": 1.3359580052493438,
+      "grad_norm": 0.4777003027087777,
+      "learning_rate": 1.1921579902562103e-05,
+      "loss": 0.6005,
+      "step": 255
+    },
+    {
+      "epoch": 1.341207349081365,
+      "grad_norm": 0.3348911333128937,
+      "learning_rate": 1.1752873650521934e-05,
+      "loss": 0.5908,
+      "step": 256
+    },
+    {
+      "epoch": 1.3464566929133859,
+      "grad_norm": 0.32219847234981497,
+      "learning_rate": 1.1584871510033707e-05,
+      "loss": 0.599,
+      "step": 257
+    },
+    {
+      "epoch": 1.3517060367454068,
+      "grad_norm": 0.32820186480956876,
+      "learning_rate": 1.1417587824547822e-05,
+      "loss": 0.5949,
+      "step": 258
+    },
+    {
+      "epoch": 1.356955380577428,
+      "grad_norm": 0.35813565851987533,
+      "learning_rate": 1.1251036876175476e-05,
+      "loss": 0.6273,
+      "step": 259
+    },
+    {
+      "epoch": 1.3622047244094488,
+      "grad_norm": 0.34029610416797407,
+      "learning_rate": 1.1085232884469236e-05,
+      "loss": 0.6026,
+      "step": 260
+    },
+    {
+      "epoch": 1.3674540682414698,
+      "grad_norm": 0.3319785571663989,
+      "learning_rate": 1.0920190005209066e-05,
+      "loss": 0.5993,
+      "step": 261
+    },
+    {
+      "epoch": 1.372703412073491,
+      "grad_norm": 0.3368980390679111,
+      "learning_rate": 1.0755922329193739e-05,
+      "loss": 0.5763,
+      "step": 262
+    },
+    {
+      "epoch": 1.3779527559055118,
+      "grad_norm": 0.32898129554439637,
+      "learning_rate": 1.0592443881037816e-05,
+      "loss": 0.5687,
+      "step": 263
+    },
+    {
+      "epoch": 1.3832020997375327,
+      "grad_norm": 0.3518104139897649,
+      "learning_rate": 1.0429768617974271e-05,
+      "loss": 0.5934,
+      "step": 264
+    },
+    {
+      "epoch": 1.3884514435695539,
+      "grad_norm": 0.329339944599712,
+      "learning_rate": 1.0267910428662878e-05,
+      "loss": 0.6384,
+      "step": 265
+    },
+    {
+      "epoch": 1.3937007874015748,
+      "grad_norm": 0.3385813426254148,
+      "learning_rate": 1.0106883132004428e-05,
+      "loss": 0.6055,
+      "step": 266
+    },
+    {
+      "epoch": 1.3989501312335957,
+      "grad_norm": 0.36895054605374145,
+      "learning_rate": 9.946700475960933e-06,
+      "loss": 0.6145,
+      "step": 267
+    },
+    {
+      "epoch": 1.4041994750656168,
+      "grad_norm": 0.3996589842241198,
+      "learning_rate": 9.787376136381866e-06,
+      "loss": 0.5953,
+      "step": 268
+    },
+    {
+      "epoch": 1.4094488188976377,
+      "grad_norm": 0.32114805305907074,
+      "learning_rate": 9.628923715836558e-06,
+      "loss": 0.5807,
+      "step": 269
+    },
+    {
+      "epoch": 1.4146981627296589,
+      "grad_norm": 0.3463735106041919,
+      "learning_rate": 9.471356742452881e-06,
+      "loss": 0.5991,
+      "step": 270
+    },
+    {
+      "epoch": 1.4199475065616798,
+      "grad_norm": 0.3218652005510028,
+      "learning_rate": 9.314688668762232e-06,
+      "loss": 0.615,
+      "step": 271
+    },
+    {
+      "epoch": 1.425196850393701,
+      "grad_norm": 0.35660358239424667,
+      "learning_rate": 9.158932870551012e-06,
+      "loss": 0.5915,
+      "step": 272
+    },
+    {
+      "epoch": 1.4304461942257218,
+      "grad_norm": 0.328623251389403,
+      "learning_rate": 9.004102645718655e-06,
+      "loss": 0.594,
+      "step": 273
+    },
+    {
+      "epoch": 1.4356955380577427,
+      "grad_norm": 0.3269276651095935,
+      "learning_rate": 8.85021121314229e-06,
+      "loss": 0.6032,
+      "step": 274
+    },
+    {
+      "epoch": 1.4409448818897639,
+      "grad_norm": 0.3563617148746268,
+      "learning_rate": 8.697271711548163e-06,
+      "loss": 0.5727,
+      "step": 275
+    },
+    {
+      "epoch": 1.4461942257217848,
+      "grad_norm": 0.3379353625795505,
+      "learning_rate": 8.545297198389896e-06,
+      "loss": 0.572,
+      "step": 276
+    },
+    {
+      "epoch": 1.4514435695538057,
+      "grad_norm": 0.32800047402633126,
+      "learning_rate": 8.394300648733688e-06,
+      "loss": 0.5784,
+      "step": 277
+    },
+    {
+      "epoch": 1.4566929133858268,
+      "grad_norm": 0.353294123585671,
+      "learning_rate": 8.24429495415054e-06,
+      "loss": 0.5948,
+      "step": 278
+    },
+    {
+      "epoch": 1.4619422572178478,
+      "grad_norm": 0.33905615123641586,
+      "learning_rate": 8.095292921615628e-06,
+      "loss": 0.6164,
+      "step": 279
+    },
+    {
+      "epoch": 1.4671916010498687,
+      "grad_norm": 0.34568179548405675,
+      "learning_rate": 7.947307272414874e-06,
+      "loss": 0.587,
+      "step": 280
+    },
+    {
+      "epoch": 1.4724409448818898,
+      "grad_norm": 0.31900836965125545,
+      "learning_rate": 7.800350641058867e-06,
+      "loss": 0.5829,
+      "step": 281
+    },
+    {
+      "epoch": 1.4776902887139107,
+      "grad_norm": 0.33258497484716926,
+      "learning_rate": 7.654435574204145e-06,
+      "loss": 0.5891,
+      "step": 282
+    },
+    {
+      "epoch": 1.4829396325459316,
+      "grad_norm": 0.32711726879832903,
+      "learning_rate": 7.509574529582022e-06,
+      "loss": 0.5915,
+      "step": 283
+    },
+    {
+      "epoch": 1.4881889763779528,
+      "grad_norm": 0.3332370756155197,
+      "learning_rate": 7.365779874934987e-06,
+      "loss": 0.5925,
+      "step": 284
+    },
+    {
+      "epoch": 1.4934383202099737,
+      "grad_norm": 0.3454414557061501,
+      "learning_rate": 7.223063886960779e-06,
+      "loss": 0.5729,
+      "step": 285
+    },
+    {
+      "epoch": 1.4986876640419948,
+      "grad_norm": 0.3332148223005662,
+      "learning_rate": 7.081438750264258e-06,
+      "loss": 0.5821,
+      "step": 286
+    },
+    {
+      "epoch": 1.5039370078740157,
+      "grad_norm": 0.36281690704994163,
+      "learning_rate": 6.940916556317119e-06,
+      "loss": 0.6177,
+      "step": 287
+    },
+    {
+      "epoch": 1.5091863517060369,
+      "grad_norm": 0.31468653111674066,
+      "learning_rate": 6.801509302425553e-06,
+      "loss": 0.5673,
+      "step": 288
+    },
+    {
+      "epoch": 1.5144356955380578,
+      "grad_norm": 0.34938019014327126,
+      "learning_rate": 6.6632288907059795e-06,
+      "loss": 0.5872,
+      "step": 289
+    },
+    {
+      "epoch": 1.5196850393700787,
+      "grad_norm": 0.34327230908091816,
+      "learning_rate": 6.526087127068857e-06,
+      "loss": 0.5667,
+      "step": 290
+    },
+    {
+      "epoch": 1.5249343832020998,
+      "grad_norm": 0.34168411959263945,
+      "learning_rate": 6.3900957202107695e-06,
+      "loss": 0.5861,
+      "step": 291
+    },
+    {
+      "epoch": 1.5301837270341208,
+      "grad_norm": 0.3336581979253816,
+      "learning_rate": 6.255266280614747e-06,
+      "loss": 0.5859,
+      "step": 292
+    },
+    {
+      "epoch": 1.5354330708661417,
+      "grad_norm": 0.3373119860626911,
+      "learning_rate": 6.1216103195590085e-06,
+      "loss": 0.6028,
+      "step": 293
+    },
+    {
+      "epoch": 1.5406824146981628,
+      "grad_norm": 0.3448522286764964,
+      "learning_rate": 5.989139248134181e-06,
+      "loss": 0.5851,
+      "step": 294
+    },
+    {
+      "epoch": 1.5459317585301837,
+      "grad_norm": 0.3186190284258791,
+      "learning_rate": 5.857864376269051e-06,
+      "loss": 0.5995,
+      "step": 295
+    },
+    {
+      "epoch": 1.5511811023622046,
+      "grad_norm": 0.6485974630036248,
+      "learning_rate": 5.727796911764955e-06,
+      "loss": 0.5983,
+      "step": 296
+    },
+    {
+      "epoch": 1.5564304461942258,
+      "grad_norm": 0.3514775336148505,
+      "learning_rate": 5.598947959338912e-06,
+      "loss": 0.5699,
+      "step": 297
+    },
+    {
+      "epoch": 1.5616797900262467,
+      "grad_norm": 0.3177218865445128,
+      "learning_rate": 5.471328519675521e-06,
+      "loss": 0.5562,
+      "step": 298
+    },
+    {
+      "epoch": 1.5669291338582676,
+      "grad_norm": 0.32933170379416266,
+      "learning_rate": 5.344949488487776e-06,
+      "loss": 0.6139,
+      "step": 299
+    },
+    {
+      "epoch": 1.5721784776902887,
+      "grad_norm": 0.5816969720018578,
+      "learning_rate": 5.219821655586821e-06,
+      "loss": 0.5993,
+      "step": 300
+    },
+    {
+      "epoch": 1.5774278215223099,
+      "grad_norm": 0.33755304547577963,
+      "learning_rate": 5.095955703960746e-06,
+      "loss": 0.6049,
+      "step": 301
+    },
+    {
+      "epoch": 1.5826771653543306,
+      "grad_norm": 0.3348322268569808,
+      "learning_rate": 4.9733622088625335e-06,
+      "loss": 0.5672,
+      "step": 302
+    },
+    {
+      "epoch": 1.5879265091863517,
+      "grad_norm": 0.31256597659310437,
+      "learning_rate": 4.852051636907144e-06,
+      "loss": 0.585,
+      "step": 303
+    },
+    {
+      "epoch": 1.5931758530183728,
+      "grad_norm": 0.3171291902225361,
+      "learning_rate": 4.732034345177941e-06,
+      "loss": 0.565,
+      "step": 304
+    },
+    {
+      "epoch": 1.5984251968503937,
+      "grad_norm": 0.3021074473107888,
+      "learning_rate": 4.613320580342422e-06,
+      "loss": 0.5658,
+      "step": 305
+    },
+    {
+      "epoch": 1.6036745406824147,
+      "grad_norm": 0.3226527803291481,
+      "learning_rate": 4.495920477777403e-06,
+      "loss": 0.565,
+      "step": 306
+    },
+    {
+      "epoch": 1.6089238845144358,
+      "grad_norm": 0.3694002496956823,
+      "learning_rate": 4.379844060703693e-06,
+      "loss": 0.5879,
+      "step": 307
+    },
+    {
+      "epoch": 1.6141732283464567,
+      "grad_norm": 0.3331352592768329,
+      "learning_rate": 4.265101239330336e-06,
+      "loss": 0.5725,
+      "step": 308
+    },
+    {
+      "epoch": 1.6194225721784776,
+      "grad_norm": 0.4522021032042819,
+      "learning_rate": 4.151701810008524e-06,
+      "loss": 0.6367,
+      "step": 309
+    },
+    {
+      "epoch": 1.6246719160104988,
+      "grad_norm": 0.33754776511498674,
+      "learning_rate": 4.03965545439521e-06,
+      "loss": 0.5991,
+      "step": 310
+    },
+    {
+      "epoch": 1.6299212598425197,
+      "grad_norm": 0.32368391205172553,
+      "learning_rate": 3.9289717386265255e-06,
+      "loss": 0.5926,
+      "step": 311
+    },
+    {
+      "epoch": 1.6351706036745406,
+      "grad_norm": 0.32475917199744364,
+      "learning_rate": 3.819660112501053e-06,
+      "loss": 0.6075,
+      "step": 312
+    },
+    {
+      "epoch": 1.6404199475065617,
+      "grad_norm": 0.3110033606981595,
+      "learning_rate": 3.711729908673034e-06,
+      "loss": 0.5888,
+      "step": 313
+    },
+    {
+      "epoch": 1.6456692913385826,
+      "grad_norm": 0.3373406386100177,
+      "learning_rate": 3.60519034185558e-06,
+      "loss": 0.5996,
+      "step": 314
+    },
+    {
+      "epoch": 1.6509186351706036,
+      "grad_norm": 0.36264975614106737,
+      "learning_rate": 3.5000505080339565e-06,
+      "loss": 0.5893,
+      "step": 315
+    },
+    {
+      "epoch": 1.6561679790026247,
+      "grad_norm": 0.3307386665641602,
+      "learning_rate": 3.3963193836889907e-06,
+      "loss": 0.6014,
+      "step": 316
+    },
+    {
+      "epoch": 1.6614173228346458,
+      "grad_norm": 0.3324854648874698,
+      "learning_rate": 3.2940058250306927e-06,
+      "loss": 0.5916,
+      "step": 317
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.32873956128719584,
+      "learning_rate": 3.193118567242148e-06,
+      "loss": 0.5894,
+      "step": 318
+    },
+    {
+      "epoch": 1.6719160104986877,
+      "grad_norm": 0.33402054955433647,
+      "learning_rate": 3.093666223733731e-06,
+      "loss": 0.578,
+      "step": 319
+    },
+    {
+      "epoch": 1.6771653543307088,
+      "grad_norm": 0.3656703893711572,
+      "learning_rate": 2.9956572854077205e-06,
+      "loss": 0.615,
+      "step": 320
+    },
+    {
+      "epoch": 1.6824146981627297,
+      "grad_norm": 0.33288296868225165,
+      "learning_rate": 2.89910011993338e-06,
+      "loss": 0.6146,
+      "step": 321
+    },
+    {
+      "epoch": 1.6876640419947506,
+      "grad_norm": 0.31270880787892713,
+      "learning_rate": 2.804002971032551e-06,
+      "loss": 0.5706,
+      "step": 322
+    },
+    {
+      "epoch": 1.6929133858267718,
+      "grad_norm": 0.36578970538344274,
+      "learning_rate": 2.7103739577758426e-06,
+      "loss": 0.5945,
+      "step": 323
+    },
+    {
+      "epoch": 1.6981627296587927,
+      "grad_norm": 0.34057921525679963,
+      "learning_rate": 2.618221073889433e-06,
+      "loss": 0.5956,
+      "step": 324
+    },
+    {
+      "epoch": 1.7034120734908136,
+      "grad_norm": 0.33694251175514184,
+      "learning_rate": 2.5275521870726107e-06,
+      "loss": 0.5859,
+      "step": 325
+    },
+    {
+      "epoch": 1.7086614173228347,
+      "grad_norm": 0.3395397908018638,
+      "learning_rate": 2.4383750383260417e-06,
+      "loss": 0.583,
+      "step": 326
+    },
+    {
+      "epoch": 1.7139107611548556,
+      "grad_norm": 0.31550755548371506,
+      "learning_rate": 2.3506972412908866e-06,
+      "loss": 0.5978,
+      "step": 327
+    },
+    {
+      "epoch": 1.7191601049868765,
+      "grad_norm": 0.33820482852297834,
+      "learning_rate": 2.264526281598762e-06,
+      "loss": 0.5794,
+      "step": 328
+    },
+    {
+      "epoch": 1.7244094488188977,
+      "grad_norm": 0.33770491265368446,
+      "learning_rate": 2.1798695162326444e-06,
+      "loss": 0.5706,
+      "step": 329
+    },
+    {
+      "epoch": 1.7296587926509186,
+      "grad_norm": 0.3289360528587284,
+      "learning_rate": 2.0967341728987554e-06,
+      "loss": 0.6085,
+      "step": 330
+    },
+    {
+      "epoch": 1.7349081364829395,
+      "grad_norm": 0.34532370867210405,
+      "learning_rate": 2.015127349409489e-06,
+      "loss": 0.5922,
+      "step": 331
+    },
+    {
+      "epoch": 1.7401574803149606,
+      "grad_norm": 0.33983136787818885,
+      "learning_rate": 1.9350560130774234e-06,
+      "loss": 0.6044,
+      "step": 332
+    },
+    {
+      "epoch": 1.7454068241469818,
+      "grad_norm": 0.324457700398482,
+      "learning_rate": 1.8565270001204693e-06,
+      "loss": 0.5918,
+      "step": 333
+    },
+    {
+      "epoch": 1.7506561679790025,
+      "grad_norm": 0.32498381242550006,
+      "learning_rate": 1.7795470150782312e-06,
+      "loss": 0.6276,
+      "step": 334
+    },
+    {
+      "epoch": 1.7559055118110236,
+      "grad_norm": 0.8612722591901686,
+      "learning_rate": 1.7041226302395797e-06,
+      "loss": 0.5829,
+      "step": 335
+    },
+    {
+      "epoch": 1.7611548556430447,
+      "grad_norm": 0.31662851542256254,
+      "learning_rate": 1.6302602850815397e-06,
+      "loss": 0.5944,
+      "step": 336
+    },
+    {
+      "epoch": 1.7664041994750657,
+      "grad_norm": 0.38212351051409194,
+      "learning_rate": 1.55796628571951e-06,
+      "loss": 0.61,
+      "step": 337
+    },
+    {
+      "epoch": 1.7716535433070866,
+      "grad_norm": 0.33186978516778165,
+      "learning_rate": 1.487246804368876e-06,
+      "loss": 0.5659,
+      "step": 338
+    },
+    {
+      "epoch": 1.7769028871391077,
+      "grad_norm": 0.3341582547520656,
+      "learning_rate": 1.418107878818027e-06,
+      "loss": 0.5944,
+      "step": 339
+    },
+    {
+      "epoch": 1.7821522309711286,
+      "grad_norm": 0.30814339482559316,
+      "learning_rate": 1.3505554119128861e-06,
+      "loss": 0.6285,
+      "step": 340
+    },
+    {
+      "epoch": 1.7874015748031495,
+      "grad_norm": 0.3386616031032492,
+      "learning_rate": 1.2845951710529513e-06,
+      "loss": 0.5887,
+      "step": 341
+    },
+    {
+      "epoch": 1.7926509186351707,
+      "grad_norm": 0.3304357792552102,
+      "learning_rate": 1.2202327876988719e-06,
+      "loss": 0.5714,
+      "step": 342
+    },
+    {
+      "epoch": 1.7979002624671916,
+      "grad_norm": 0.3216612982270511,
+      "learning_rate": 1.157473756891674e-06,
+      "loss": 0.5886,
+      "step": 343
+    },
+    {
+      "epoch": 1.8031496062992125,
+      "grad_norm": 0.32347441677351674,
+      "learning_rate": 1.0963234367836106e-06,
+      "loss": 0.5773,
+      "step": 344
+    },
+    {
+      "epoch": 1.8083989501312336,
+      "grad_norm": 0.3259998522864815,
+      "learning_rate": 1.036787048180683e-06,
+      "loss": 0.6001,
+      "step": 345
+    },
+    {
+      "epoch": 1.8136482939632546,
+      "grad_norm": 0.3283222158898781,
+      "learning_rate": 9.788696740969295e-07,
+      "loss": 0.5579,
+      "step": 346
+    },
+    {
+      "epoch": 1.8188976377952755,
+      "grad_norm": 0.3306253571853781,
+      "learning_rate": 9.225762593204379e-07,
+      "loss": 0.6226,
+      "step": 347
+    },
+    {
+      "epoch": 1.8241469816272966,
+      "grad_norm": 0.3152113248795258,
+      "learning_rate": 8.679116099911855e-07,
+      "loss": 0.6108,
+      "step": 348
+    },
+    {
+      "epoch": 1.8293963254593177,
+      "grad_norm": 0.3402669579215478,
+      "learning_rate": 8.148803931907023e-07,
+      "loss": 0.5892,
+      "step": 349
+    },
+    {
+      "epoch": 1.8346456692913384,
+      "grad_norm": 0.34444338413492365,
+      "learning_rate": 7.634871365436192e-07,
+      "loss": 0.5899,
+      "step": 350
+    },
+    {
+      "epoch": 1.8398950131233596,
+      "grad_norm": 0.36262207426472215,
+      "learning_rate": 7.137362278311033e-07,
+      "loss": 0.6489,
+      "step": 351
+    },
+    {
+      "epoch": 1.8451443569553807,
+      "grad_norm": 0.3153477780162296,
+      "learning_rate": 6.656319146162516e-07,
+      "loss": 0.5755,
+      "step": 352
+    },
+    {
+      "epoch": 1.8503937007874016,
+      "grad_norm": 0.3341125391990547,
+      "learning_rate": 6.191783038814492e-07,
+      "loss": 0.5997,
+      "step": 353
+    },
+    {
+      "epoch": 1.8556430446194225,
+      "grad_norm": 0.3048954224460625,
+      "learning_rate": 5.743793616777216e-07,
+      "loss": 0.5793,
+      "step": 354
+    },
+    {
+      "epoch": 1.8608923884514437,
+      "grad_norm": 0.3246033177970451,
+      "learning_rate": 5.312389127861428e-07,
+      "loss": 0.5687,
+      "step": 355
+    },
+    {
+      "epoch": 1.8661417322834646,
+      "grad_norm": 0.3137655487999818,
+      "learning_rate": 4.89760640391268e-07,
+      "loss": 0.6075,
+      "step": 356
+    },
+    {
+      "epoch": 1.8713910761154855,
+      "grad_norm": 0.34369114683688085,
+      "learning_rate": 4.499480857666849e-07,
+      "loss": 0.5838,
+      "step": 357
+    },
+    {
+      "epoch": 1.8766404199475066,
+      "grad_norm": 0.3217429356827238,
+      "learning_rate": 4.118046479726823e-07,
+      "loss": 0.5598,
+      "step": 358
+    },
+    {
+      "epoch": 1.8818897637795275,
+      "grad_norm": 0.3002853847320703,
+      "learning_rate": 3.75333583566031e-07,
+      "loss": 0.5991,
+      "step": 359
+    },
+    {
+      "epoch": 1.8871391076115485,
+      "grad_norm": 0.3369774600083623,
+      "learning_rate": 3.4053800632196434e-07,
+      "loss": 0.5631,
+      "step": 360
+    },
+    {
+      "epoch": 1.8923884514435696,
+      "grad_norm": 0.30143355641339575,
+      "learning_rate": 3.074208869683282e-07,
+      "loss": 0.5653,
+      "step": 361
+    },
+    {
+      "epoch": 1.8976377952755905,
+      "grad_norm": 0.36214676043014177,
+      "learning_rate": 2.7598505293194855e-07,
+      "loss": 0.6766,
+      "step": 362
+    },
+    {
+      "epoch": 1.9028871391076114,
+      "grad_norm": 0.2987940805703468,
+      "learning_rate": 2.462331880972468e-07,
+      "loss": 0.5643,
+      "step": 363
+    },
+    {
+      "epoch": 1.9081364829396326,
+      "grad_norm": 0.320306518159664,
+      "learning_rate": 2.1816783257708084e-07,
+      "loss": 0.5736,
+      "step": 364
+    },
+    {
+      "epoch": 1.9133858267716537,
+      "grad_norm": 0.3206998817539087,
+      "learning_rate": 1.9179138249589836e-07,
+      "loss": 0.5697,
+      "step": 365
+    },
+    {
+      "epoch": 1.9186351706036744,
+      "grad_norm": 0.32465002287627187,
+      "learning_rate": 1.6710608978514509e-07,
+      "loss": 0.6046,
+      "step": 366
+    },
+    {
+      "epoch": 1.9238845144356955,
+      "grad_norm": 0.3277788035058438,
+      "learning_rate": 1.4411406199102084e-07,
+      "loss": 0.604,
+      "step": 367
+    },
+    {
+      "epoch": 1.9291338582677167,
+      "grad_norm": 0.33068955282697027,
+      "learning_rate": 1.2281726209452782e-07,
+      "loss": 0.5839,
+      "step": 368
+    },
+    {
+      "epoch": 1.9343832020997376,
+      "grad_norm": 0.34218514874858724,
+      "learning_rate": 1.0321750834388911e-07,
+      "loss": 0.585,
+      "step": 369
+    },
+    {
+      "epoch": 1.9396325459317585,
+      "grad_norm": 0.31844253299268915,
+      "learning_rate": 8.531647409931065e-08,
+      "loss": 0.5639,
+      "step": 370
+    },
+    {
+      "epoch": 1.9448818897637796,
+      "grad_norm": 0.30830039529015196,
+      "learning_rate": 6.91156876901089e-08,
+      "loss": 0.6093,
+      "step": 371
+    },
+    {
+      "epoch": 1.9501312335958005,
+      "grad_norm": 0.32498682800186096,
+      "learning_rate": 5.4616532284239576e-08,
+      "loss": 0.562,
+      "step": 372
+    },
+    {
+      "epoch": 1.9553805774278215,
+      "grad_norm": 0.3092071908083608,
+      "learning_rate": 4.182024577019439e-08,
+      "loss": 0.5962,
+      "step": 373
+    },
+    {
+      "epoch": 1.9606299212598426,
+      "grad_norm": 0.5172318996037643,
+      "learning_rate": 3.072792065132113e-08,
+      "loss": 0.604,
+      "step": 374
+    },
+    {
+      "epoch": 1.9658792650918635,
+      "grad_norm": 0.32068020240293016,
+      "learning_rate": 2.1340503952551606e-08,
+      "loss": 0.5956,
+      "step": 375
+    },
+    {
+      "epoch": 1.9711286089238844,
+      "grad_norm": 0.327509114877014,
+      "learning_rate": 1.365879713954188e-08,
+      "loss": 0.5611,
+      "step": 376
+    },
+    {
+      "epoch": 1.9763779527559056,
+      "grad_norm": 0.3419187964583904,
+      "learning_rate": 7.683456050251447e-09,
+      "loss": 0.5885,
+      "step": 377
+    },
+    {
+      "epoch": 1.9816272965879265,
+      "grad_norm": 0.32624555447658893,
+      "learning_rate": 3.414990838945809e-09,
+      "loss": 0.5895,
+      "step": 378
+    },
+    {
+      "epoch": 1.9868766404199474,
+      "grad_norm": 0.3400907937056279,
+      "learning_rate": 8.537659326424141e-10,
+      "loss": 0.5784,
+      "step": 379
+    },
+    {
+      "epoch": 1.9921259842519685,
+      "grad_norm": 0.32480553293831377,
+      "learning_rate": 0.0,
+      "loss": 0.5703,
+      "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 sha256:6c0d2528dcfd8d62d3c517248c2d231cc9ff64ec148911ec3ce58a9d39f7507d
+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)