Training in progress, step 2200, checkpoint

.gitattributes

@@ -37,3 +37,4 @@ tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoint-550/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoint-1100/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoint-1650/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-2200/tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoint-2200/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|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
+}

checkpoint-2200/config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "ZMC2019/Qwen2.5-Math-7B-Instruct",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 300000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

checkpoint-2200/generation_config.json

@@ -0,0 +1,9 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "transformers_version": "4.49.0"
+}

checkpoint-2200/global_step2200/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36fe7a20d10e5d90106430a268fad92466c60a921af6b5316a61faf64d5be6d5
+size 11423429708

checkpoint-2200/global_step2200/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f2ccccb51734e5487646c471a342297220c748b72156af50c40620c99b54700
+size 11423429708

checkpoint-2200/global_step2200/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:491f100f4f0501afddfe9c978d651d0c3425d244dde56595bc7839c0fc8052a0
+size 11423429708

checkpoint-2200/global_step2200/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7ab9e75d56c5826e0de7845f1ad42875ae69519977f52a970cc4985ece3d60b
+size 11423429708

checkpoint-2200/global_step2200/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be166f92d224f7073108d200ac000a5104f59c9ba44865894c035ec3c9631fa9
+size 11423429708

checkpoint-2200/global_step2200/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce154f803343369023169e297ab22f4c784cb4e78c8213e70745e5572ed768b9
+size 11423429708

checkpoint-2200/global_step2200/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14f32f9f7cd3686e73c6605100dedbdb71d55abe74dbaaec843098ffa610401f
+size 11423429708

checkpoint-2200/global_step2200/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b28e14a149912b0c7cb42dba554bf624157d841ca670a1d1c1b7c983f8e9b8c
+size 11423429708

checkpoint-2200/global_step2200/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:877e520477bc3424104fdfc8dad8d5d6e6f94a40cc1e614e4a64fcc4f5864f9e
+size 164885

checkpoint-2200/global_step2200/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4887622e782f6eefd68430d06c6fc4b4a4f18305550a2245e3817890e5c1aa7
+size 164885

checkpoint-2200/global_step2200/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:860bf99fdf7d23296dafdd799612c49d5e0122b35cbca8ea76c917beefbfa522
+size 164885

checkpoint-2200/global_step2200/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e12d39794c930d442508d30d8cf95543165fe8c8917de477c53cd4d013dcd958
+size 164885

checkpoint-2200/global_step2200/zero_pp_rank_4_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9e7c65965f129119059c6a57c7a7322ee09f907ff35ecbc959a6205142250c5
+size 164885

checkpoint-2200/global_step2200/zero_pp_rank_5_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9709d2d8ad876889faa375c9b6ff7270e8408f8642b78450808da21ff85925ca
+size 164885

checkpoint-2200/global_step2200/zero_pp_rank_6_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:186d7977377b82a9b0fb9c1b1130db0649a8cc975375377bf592ab5249768b62
+size 164885

checkpoint-2200/global_step2200/zero_pp_rank_7_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb25b8a9335cee877843153483a761fcb5458681a9f26f2ccbd41b11c443e71e
+size 164885

checkpoint-2200/latest

@@ -0,0 +1 @@
+global_step2200

checkpoint-2200/model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d78f6c6e06cb7bfd318baf401c4efa8d2f0a2f325c53b1f9b34512cc2f756d96
+size 4877660776

checkpoint-2200/model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce864a76eab9bbe8dd349a9108c7355bf7a92f19452b1f936cc720b6e78ab468
+size 4932751008

checkpoint-2200/model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3fded280965a9697cb46f9f287e972aab63c23a82226e64666f14edf811d4c3
+size 4330865200

checkpoint-2200/model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2acc55d626073b9ef2a7a6382a2ece9b0c090c6f386f8ca101fe2e94cfa7e962
+size 1089994880

checkpoint-2200/model.safetensors.index.json

@@ -0,0 +1,346 @@
+{
+  "metadata": {
+    "total_size": 15231233024
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00003-of-00004.safetensors"
+  }
+}

checkpoint-2200/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08282b46825aa78d10fe10e3fea89555c5b5a691b261a3ddfd58fcb58370edff
+size 15984

checkpoint-2200/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbab71d98a3a9a92df82a6bba463947327c3a1bcf35cd9f4f46114641fc42dd9
+size 15984

checkpoint-2200/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:caac82d57d878d30219a4f9ec289a97ff90c53afc160b968f251b3fd3454b8d8
+size 15984

checkpoint-2200/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19762d2d370222b01817da11bbaa6665d542293373186d66f754e7246bb861ed
+size 15984

checkpoint-2200/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00c7508b346a7d3c5c23392845f1d013331114ade778794b76e919cb3ed5d33e
+size 15984

checkpoint-2200/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b89de7d14dd20a191f56b74c816ef8b7fe5c171e31efbeadbf321c4539ed68c3
+size 15984

checkpoint-2200/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c71152053553e6e22d670fbc4fd7550bf8a046b54cad7b71869787986a6a42c
+size 15984

checkpoint-2200/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b67db12a26a26ffe03d9afc84a43857eb2e5b2fec2dd189653b415f74208190
+size 15984

checkpoint-2200/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54a6219d359c6fef1cbfa3f55a59595228ae0b1d09223b5d474db3a6957c1d42
+size 1064

checkpoint-2200/special_tokens_map.json

@@ -0,0 +1,25 @@
+{
+  "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": "<|im_end|>"
+}

checkpoint-2200/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

checkpoint-2200/tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "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
+    }
+  },
+  "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        {{- 'Please reason step by step, and put your final answer within \\\\boxed{}.' }}\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    {%- else %}\n        {{- '<|im_start|>system\\nPlease reason step by step, and put your final answer within \\\\boxed{}.<|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) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.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' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-2200/trainer_state.json

@@ -0,0 +1,3113 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.050326188257223,
+  "eval_steps": 500,
+  "global_step": 2200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.004659832246039142,
+      "grad_norm": 55.709484937361424,
+      "learning_rate": 7.763975155279503e-07,
+      "loss": 11.0694,
+      "step": 5
+    },
+    {
+      "epoch": 0.009319664492078284,
+      "grad_norm": 57.1398137451701,
+      "learning_rate": 1.5527950310559006e-06,
+      "loss": 10.8846,
+      "step": 10
+    },
+    {
+      "epoch": 0.013979496738117428,
+      "grad_norm": 97.98047478379912,
+      "learning_rate": 2.329192546583851e-06,
+      "loss": 9.4413,
+      "step": 15
+    },
+    {
+      "epoch": 0.01863932898415657,
+      "grad_norm": 32.76887701215898,
+      "learning_rate": 3.1055900621118013e-06,
+      "loss": 3.214,
+      "step": 20
+    },
+    {
+      "epoch": 0.023299161230195712,
+      "grad_norm": 3.315394944010228,
+      "learning_rate": 3.881987577639752e-06,
+      "loss": 1.346,
+      "step": 25
+    },
+    {
+      "epoch": 0.027958993476234855,
+      "grad_norm": 1.29231664681427,
+      "learning_rate": 4.658385093167702e-06,
+      "loss": 1.006,
+      "step": 30
+    },
+    {
+      "epoch": 0.032618825722273995,
+      "grad_norm": 0.8629586527731651,
+      "learning_rate": 5.4347826086956525e-06,
+      "loss": 0.8411,
+      "step": 35
+    },
+    {
+      "epoch": 0.03727865796831314,
+      "grad_norm": 0.6212648141101664,
+      "learning_rate": 6.2111801242236025e-06,
+      "loss": 0.7754,
+      "step": 40
+    },
+    {
+      "epoch": 0.04193849021435228,
+      "grad_norm": 0.46040470834395936,
+      "learning_rate": 6.9875776397515525e-06,
+      "loss": 0.7136,
+      "step": 45
+    },
+    {
+      "epoch": 0.046598322460391424,
+      "grad_norm": 0.42693871288027446,
+      "learning_rate": 7.763975155279503e-06,
+      "loss": 0.6831,
+      "step": 50
+    },
+    {
+      "epoch": 0.05125815470643057,
+      "grad_norm": 0.44192720551532577,
+      "learning_rate": 8.540372670807453e-06,
+      "loss": 0.646,
+      "step": 55
+    },
+    {
+      "epoch": 0.05591798695246971,
+      "grad_norm": 0.31222059058105195,
+      "learning_rate": 9.316770186335403e-06,
+      "loss": 0.6088,
+      "step": 60
+    },
+    {
+      "epoch": 0.06057781919850885,
+      "grad_norm": 0.30027349299916717,
+      "learning_rate": 1.0093167701863353e-05,
+      "loss": 0.5932,
+      "step": 65
+    },
+    {
+      "epoch": 0.06523765144454799,
+      "grad_norm": 0.3320480978786744,
+      "learning_rate": 1.0869565217391305e-05,
+      "loss": 0.5892,
+      "step": 70
+    },
+    {
+      "epoch": 0.06989748369058714,
+      "grad_norm": 0.3001101299512147,
+      "learning_rate": 1.1645962732919255e-05,
+      "loss": 0.5713,
+      "step": 75
+    },
+    {
+      "epoch": 0.07455731593662628,
+      "grad_norm": 0.30892715626333545,
+      "learning_rate": 1.2422360248447205e-05,
+      "loss": 0.5619,
+      "step": 80
+    },
+    {
+      "epoch": 0.07921714818266543,
+      "grad_norm": 0.3319072685834915,
+      "learning_rate": 1.3198757763975155e-05,
+      "loss": 0.5525,
+      "step": 85
+    },
+    {
+      "epoch": 0.08387698042870456,
+      "grad_norm": 0.29047402255141175,
+      "learning_rate": 1.3975155279503105e-05,
+      "loss": 0.5628,
+      "step": 90
+    },
+    {
+      "epoch": 0.08853681267474371,
+      "grad_norm": 0.25922715271689156,
+      "learning_rate": 1.4751552795031057e-05,
+      "loss": 0.5443,
+      "step": 95
+    },
+    {
+      "epoch": 0.09319664492078285,
+      "grad_norm": 0.31531185304643106,
+      "learning_rate": 1.5527950310559007e-05,
+      "loss": 0.5296,
+      "step": 100
+    },
+    {
+      "epoch": 0.097856477166822,
+      "grad_norm": 0.3013647908850556,
+      "learning_rate": 1.630434782608696e-05,
+      "loss": 0.5243,
+      "step": 105
+    },
+    {
+      "epoch": 0.10251630941286113,
+      "grad_norm": 0.3047350591065195,
+      "learning_rate": 1.7080745341614907e-05,
+      "loss": 0.5261,
+      "step": 110
+    },
+    {
+      "epoch": 0.10717614165890028,
+      "grad_norm": 0.30780196150291267,
+      "learning_rate": 1.785714285714286e-05,
+      "loss": 0.5231,
+      "step": 115
+    },
+    {
+      "epoch": 0.11183597390493942,
+      "grad_norm": 0.3804666896095633,
+      "learning_rate": 1.8633540372670807e-05,
+      "loss": 0.5098,
+      "step": 120
+    },
+    {
+      "epoch": 0.11649580615097857,
+      "grad_norm": 0.2868238329098672,
+      "learning_rate": 1.940993788819876e-05,
+      "loss": 0.5057,
+      "step": 125
+    },
+    {
+      "epoch": 0.1211556383970177,
+      "grad_norm": 0.32341698454060924,
+      "learning_rate": 2.0186335403726707e-05,
+      "loss": 0.5114,
+      "step": 130
+    },
+    {
+      "epoch": 0.12581547064305684,
+      "grad_norm": 0.4221000244611152,
+      "learning_rate": 2.096273291925466e-05,
+      "loss": 0.5106,
+      "step": 135
+    },
+    {
+      "epoch": 0.13047530288909598,
+      "grad_norm": 0.47197386491626697,
+      "learning_rate": 2.173913043478261e-05,
+      "loss": 0.4951,
+      "step": 140
+    },
+    {
+      "epoch": 0.13513513513513514,
+      "grad_norm": 0.3721540533945164,
+      "learning_rate": 2.2515527950310562e-05,
+      "loss": 0.522,
+      "step": 145
+    },
+    {
+      "epoch": 0.13979496738117428,
+      "grad_norm": 0.4388944569421825,
+      "learning_rate": 2.329192546583851e-05,
+      "loss": 0.4959,
+      "step": 150
+    },
+    {
+      "epoch": 0.14445479962721341,
+      "grad_norm": 0.4622898334441083,
+      "learning_rate": 2.4068322981366462e-05,
+      "loss": 0.5041,
+      "step": 155
+    },
+    {
+      "epoch": 0.14911463187325255,
+      "grad_norm": 0.49630920093444675,
+      "learning_rate": 2.484472049689441e-05,
+      "loss": 0.4849,
+      "step": 160
+    },
+    {
+      "epoch": 0.15377446411929171,
+      "grad_norm": 0.46849210325540636,
+      "learning_rate": 2.5621118012422362e-05,
+      "loss": 0.4961,
+      "step": 165
+    },
+    {
+      "epoch": 0.15843429636533085,
+      "grad_norm": 0.4082563546888238,
+      "learning_rate": 2.639751552795031e-05,
+      "loss": 0.4719,
+      "step": 170
+    },
+    {
+      "epoch": 0.16309412861137,
+      "grad_norm": 0.4549138940471945,
+      "learning_rate": 2.7173913043478262e-05,
+      "loss": 0.4852,
+      "step": 175
+    },
+    {
+      "epoch": 0.16775396085740912,
+      "grad_norm": 0.41032309538662626,
+      "learning_rate": 2.795031055900621e-05,
+      "loss": 0.478,
+      "step": 180
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 0.36866874054662224,
+      "learning_rate": 2.8726708074534165e-05,
+      "loss": 0.481,
+      "step": 185
+    },
+    {
+      "epoch": 0.17707362534948742,
+      "grad_norm": 0.36280659725390446,
+      "learning_rate": 2.9503105590062114e-05,
+      "loss": 0.4908,
+      "step": 190
+    },
+    {
+      "epoch": 0.18173345759552656,
+      "grad_norm": 0.4446282244198985,
+      "learning_rate": 3.0279503105590062e-05,
+      "loss": 0.4802,
+      "step": 195
+    },
+    {
+      "epoch": 0.1863932898415657,
+      "grad_norm": 0.4818264381472692,
+      "learning_rate": 3.1055900621118014e-05,
+      "loss": 0.4818,
+      "step": 200
+    },
+    {
+      "epoch": 0.19105312208760486,
+      "grad_norm": 0.5191372230429591,
+      "learning_rate": 3.183229813664597e-05,
+      "loss": 0.4708,
+      "step": 205
+    },
+    {
+      "epoch": 0.195712954333644,
+      "grad_norm": 0.5690145277815086,
+      "learning_rate": 3.260869565217392e-05,
+      "loss": 0.4848,
+      "step": 210
+    },
+    {
+      "epoch": 0.20037278657968313,
+      "grad_norm": 0.41006685540485915,
+      "learning_rate": 3.3385093167701865e-05,
+      "loss": 0.467,
+      "step": 215
+    },
+    {
+      "epoch": 0.20503261882572227,
+      "grad_norm": 0.47272355547942274,
+      "learning_rate": 3.4161490683229814e-05,
+      "loss": 0.4716,
+      "step": 220
+    },
+    {
+      "epoch": 0.2096924510717614,
+      "grad_norm": 0.5135260711958035,
+      "learning_rate": 3.493788819875777e-05,
+      "loss": 0.4536,
+      "step": 225
+    },
+    {
+      "epoch": 0.21435228331780057,
+      "grad_norm": 0.5124408748792798,
+      "learning_rate": 3.571428571428572e-05,
+      "loss": 0.467,
+      "step": 230
+    },
+    {
+      "epoch": 0.2190121155638397,
+      "grad_norm": 0.4205466417850958,
+      "learning_rate": 3.6490683229813665e-05,
+      "loss": 0.4647,
+      "step": 235
+    },
+    {
+      "epoch": 0.22367194780987884,
+      "grad_norm": 0.6251611138936444,
+      "learning_rate": 3.7267080745341614e-05,
+      "loss": 0.4624,
+      "step": 240
+    },
+    {
+      "epoch": 0.22833178005591798,
+      "grad_norm": 0.5610747747609784,
+      "learning_rate": 3.804347826086957e-05,
+      "loss": 0.4676,
+      "step": 245
+    },
+    {
+      "epoch": 0.23299161230195714,
+      "grad_norm": 0.4692752738735764,
+      "learning_rate": 3.881987577639752e-05,
+      "loss": 0.4592,
+      "step": 250
+    },
+    {
+      "epoch": 0.23765144454799628,
+      "grad_norm": 0.5435654364057863,
+      "learning_rate": 3.9596273291925465e-05,
+      "loss": 0.4679,
+      "step": 255
+    },
+    {
+      "epoch": 0.2423112767940354,
+      "grad_norm": 0.5264727485653637,
+      "learning_rate": 4.0372670807453414e-05,
+      "loss": 0.463,
+      "step": 260
+    },
+    {
+      "epoch": 0.24697110904007455,
+      "grad_norm": 0.5096976732033014,
+      "learning_rate": 4.114906832298137e-05,
+      "loss": 0.4613,
+      "step": 265
+    },
+    {
+      "epoch": 0.2516309412861137,
+      "grad_norm": 0.4649125930738299,
+      "learning_rate": 4.192546583850932e-05,
+      "loss": 0.4594,
+      "step": 270
+    },
+    {
+      "epoch": 0.25629077353215285,
+      "grad_norm": 0.5173502447936299,
+      "learning_rate": 4.270186335403727e-05,
+      "loss": 0.4652,
+      "step": 275
+    },
+    {
+      "epoch": 0.26095060577819196,
+      "grad_norm": 0.3868002746815184,
+      "learning_rate": 4.347826086956522e-05,
+      "loss": 0.4572,
+      "step": 280
+    },
+    {
+      "epoch": 0.2656104380242311,
+      "grad_norm": 0.5001940885551612,
+      "learning_rate": 4.425465838509317e-05,
+      "loss": 0.4613,
+      "step": 285
+    },
+    {
+      "epoch": 0.2702702702702703,
+      "grad_norm": 0.6213239402228385,
+      "learning_rate": 4.5031055900621124e-05,
+      "loss": 0.4633,
+      "step": 290
+    },
+    {
+      "epoch": 0.2749301025163094,
+      "grad_norm": 0.6125652570862165,
+      "learning_rate": 4.580745341614907e-05,
+      "loss": 0.4673,
+      "step": 295
+    },
+    {
+      "epoch": 0.27958993476234856,
+      "grad_norm": 0.7479197761570839,
+      "learning_rate": 4.658385093167702e-05,
+      "loss": 0.4693,
+      "step": 300
+    },
+    {
+      "epoch": 0.2842497670083877,
+      "grad_norm": 0.5483826450591869,
+      "learning_rate": 4.736024844720497e-05,
+      "loss": 0.4579,
+      "step": 305
+    },
+    {
+      "epoch": 0.28890959925442683,
+      "grad_norm": 0.6644586513259841,
+      "learning_rate": 4.8136645962732924e-05,
+      "loss": 0.45,
+      "step": 310
+    },
+    {
+      "epoch": 0.293569431500466,
+      "grad_norm": 0.8188490252541196,
+      "learning_rate": 4.891304347826087e-05,
+      "loss": 0.459,
+      "step": 315
+    },
+    {
+      "epoch": 0.2982292637465051,
+      "grad_norm": 1.034709523419619,
+      "learning_rate": 4.968944099378882e-05,
+      "loss": 0.4441,
+      "step": 320
+    },
+    {
+      "epoch": 0.30288909599254427,
+      "grad_norm": 0.759919110757548,
+      "learning_rate": 4.994822229892993e-05,
+      "loss": 0.4415,
+      "step": 325
+    },
+    {
+      "epoch": 0.30754892823858343,
+      "grad_norm": 0.6470711913841568,
+      "learning_rate": 4.986192613047981e-05,
+      "loss": 0.4519,
+      "step": 330
+    },
+    {
+      "epoch": 0.31220876048462254,
+      "grad_norm": 0.8153464642106912,
+      "learning_rate": 4.977562996202969e-05,
+      "loss": 0.4527,
+      "step": 335
+    },
+    {
+      "epoch": 0.3168685927306617,
+      "grad_norm": 0.4698984383111023,
+      "learning_rate": 4.968933379357957e-05,
+      "loss": 0.4532,
+      "step": 340
+    },
+    {
+      "epoch": 0.32152842497670087,
+      "grad_norm": 0.4756319068347418,
+      "learning_rate": 4.9603037625129445e-05,
+      "loss": 0.4479,
+      "step": 345
+    },
+    {
+      "epoch": 0.32618825722274,
+      "grad_norm": 0.35969300915800667,
+      "learning_rate": 4.951674145667933e-05,
+      "loss": 0.4613,
+      "step": 350
+    },
+    {
+      "epoch": 0.33084808946877914,
+      "grad_norm": 0.5175115868112082,
+      "learning_rate": 4.94304452882292e-05,
+      "loss": 0.4473,
+      "step": 355
+    },
+    {
+      "epoch": 0.33550792171481825,
+      "grad_norm": 0.46094013566363407,
+      "learning_rate": 4.934414911977908e-05,
+      "loss": 0.445,
+      "step": 360
+    },
+    {
+      "epoch": 0.3401677539608574,
+      "grad_norm": 0.4104764392290741,
+      "learning_rate": 4.9257852951328965e-05,
+      "loss": 0.4415,
+      "step": 365
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 0.40096213441032663,
+      "learning_rate": 4.917155678287884e-05,
+      "loss": 0.4546,
+      "step": 370
+    },
+    {
+      "epoch": 0.3494874184529357,
+      "grad_norm": 0.4645867033422015,
+      "learning_rate": 4.908526061442872e-05,
+      "loss": 0.4408,
+      "step": 375
+    },
+    {
+      "epoch": 0.35414725069897485,
+      "grad_norm": 0.4400737008116359,
+      "learning_rate": 4.89989644459786e-05,
+      "loss": 0.4557,
+      "step": 380
+    },
+    {
+      "epoch": 0.35880708294501396,
+      "grad_norm": 0.4260635316927096,
+      "learning_rate": 4.891266827752848e-05,
+      "loss": 0.4575,
+      "step": 385
+    },
+    {
+      "epoch": 0.3634669151910531,
+      "grad_norm": 0.5323727061907297,
+      "learning_rate": 4.882637210907836e-05,
+      "loss": 0.4417,
+      "step": 390
+    },
+    {
+      "epoch": 0.3681267474370923,
+      "grad_norm": 0.5836242210222355,
+      "learning_rate": 4.874007594062824e-05,
+      "loss": 0.4481,
+      "step": 395
+    },
+    {
+      "epoch": 0.3727865796831314,
+      "grad_norm": 0.4273563357499149,
+      "learning_rate": 4.865377977217811e-05,
+      "loss": 0.4327,
+      "step": 400
+    },
+    {
+      "epoch": 0.37744641192917056,
+      "grad_norm": 0.3953667888190435,
+      "learning_rate": 4.8567483603728e-05,
+      "loss": 0.4512,
+      "step": 405
+    },
+    {
+      "epoch": 0.3821062441752097,
+      "grad_norm": 0.41871105603942904,
+      "learning_rate": 4.8481187435277875e-05,
+      "loss": 0.4471,
+      "step": 410
+    },
+    {
+      "epoch": 0.38676607642124883,
+      "grad_norm": 0.5437922609735028,
+      "learning_rate": 4.839489126682776e-05,
+      "loss": 0.4338,
+      "step": 415
+    },
+    {
+      "epoch": 0.391425908667288,
+      "grad_norm": 0.5564237695130677,
+      "learning_rate": 4.830859509837763e-05,
+      "loss": 0.4521,
+      "step": 420
+    },
+    {
+      "epoch": 0.3960857409133271,
+      "grad_norm": 0.34206255911684263,
+      "learning_rate": 4.822229892992751e-05,
+      "loss": 0.4495,
+      "step": 425
+    },
+    {
+      "epoch": 0.40074557315936626,
+      "grad_norm": 0.4019378807561449,
+      "learning_rate": 4.8136002761477395e-05,
+      "loss": 0.4365,
+      "step": 430
+    },
+    {
+      "epoch": 0.40540540540540543,
+      "grad_norm": 0.5440015382583567,
+      "learning_rate": 4.804970659302727e-05,
+      "loss": 0.4499,
+      "step": 435
+    },
+    {
+      "epoch": 0.41006523765144454,
+      "grad_norm": 0.5537615079643711,
+      "learning_rate": 4.796341042457715e-05,
+      "loss": 0.4438,
+      "step": 440
+    },
+    {
+      "epoch": 0.4147250698974837,
+      "grad_norm": 0.586960808161654,
+      "learning_rate": 4.787711425612703e-05,
+      "loss": 0.4386,
+      "step": 445
+    },
+    {
+      "epoch": 0.4193849021435228,
+      "grad_norm": 0.5072343488418793,
+      "learning_rate": 4.779081808767691e-05,
+      "loss": 0.4292,
+      "step": 450
+    },
+    {
+      "epoch": 0.424044734389562,
+      "grad_norm": 0.48278977249280486,
+      "learning_rate": 4.770452191922679e-05,
+      "loss": 0.4362,
+      "step": 455
+    },
+    {
+      "epoch": 0.42870456663560114,
+      "grad_norm": 0.39818179591075503,
+      "learning_rate": 4.761822575077667e-05,
+      "loss": 0.4329,
+      "step": 460
+    },
+    {
+      "epoch": 0.43336439888164024,
+      "grad_norm": 0.48993775783018667,
+      "learning_rate": 4.753192958232654e-05,
+      "loss": 0.4299,
+      "step": 465
+    },
+    {
+      "epoch": 0.4380242311276794,
+      "grad_norm": 0.41171920657337485,
+      "learning_rate": 4.744563341387643e-05,
+      "loss": 0.4428,
+      "step": 470
+    },
+    {
+      "epoch": 0.4426840633737186,
+      "grad_norm": 0.4400559670696649,
+      "learning_rate": 4.7359337245426306e-05,
+      "loss": 0.432,
+      "step": 475
+    },
+    {
+      "epoch": 0.4473438956197577,
+      "grad_norm": 0.4686484619433322,
+      "learning_rate": 4.7273041076976184e-05,
+      "loss": 0.4441,
+      "step": 480
+    },
+    {
+      "epoch": 0.45200372786579684,
+      "grad_norm": 0.4455986109270811,
+      "learning_rate": 4.718674490852606e-05,
+      "loss": 0.4368,
+      "step": 485
+    },
+    {
+      "epoch": 0.45666356011183595,
+      "grad_norm": 0.4366809315888367,
+      "learning_rate": 4.710044874007594e-05,
+      "loss": 0.4312,
+      "step": 490
+    },
+    {
+      "epoch": 0.4613233923578751,
+      "grad_norm": 0.4587131719176713,
+      "learning_rate": 4.7014152571625826e-05,
+      "loss": 0.4369,
+      "step": 495
+    },
+    {
+      "epoch": 0.4659832246039143,
+      "grad_norm": 0.5269402211397143,
+      "learning_rate": 4.6927856403175704e-05,
+      "loss": 0.44,
+      "step": 500
+    },
+    {
+      "epoch": 0.4706430568499534,
+      "grad_norm": 0.387320399512531,
+      "learning_rate": 4.684156023472558e-05,
+      "loss": 0.4266,
+      "step": 505
+    },
+    {
+      "epoch": 0.47530288909599255,
+      "grad_norm": 0.4446811703899567,
+      "learning_rate": 4.675526406627546e-05,
+      "loss": 0.435,
+      "step": 510
+    },
+    {
+      "epoch": 0.47996272134203166,
+      "grad_norm": 0.6228000644001809,
+      "learning_rate": 4.666896789782534e-05,
+      "loss": 0.433,
+      "step": 515
+    },
+    {
+      "epoch": 0.4846225535880708,
+      "grad_norm": 0.5372584652001571,
+      "learning_rate": 4.658267172937522e-05,
+      "loss": 0.4368,
+      "step": 520
+    },
+    {
+      "epoch": 0.48928238583411,
+      "grad_norm": 0.4279339582606534,
+      "learning_rate": 4.64963755609251e-05,
+      "loss": 0.4358,
+      "step": 525
+    },
+    {
+      "epoch": 0.4939422180801491,
+      "grad_norm": 0.5212302132415493,
+      "learning_rate": 4.641007939247497e-05,
+      "loss": 0.4327,
+      "step": 530
+    },
+    {
+      "epoch": 0.49860205032618826,
+      "grad_norm": 0.38993326302525777,
+      "learning_rate": 4.632378322402486e-05,
+      "loss": 0.4329,
+      "step": 535
+    },
+    {
+      "epoch": 0.5032618825722274,
+      "grad_norm": 0.3927986683558084,
+      "learning_rate": 4.6237487055574736e-05,
+      "loss": 0.4228,
+      "step": 540
+    },
+    {
+      "epoch": 0.5079217148182665,
+      "grad_norm": 0.48245624377171165,
+      "learning_rate": 4.6151190887124615e-05,
+      "loss": 0.4292,
+      "step": 545
+    },
+    {
+      "epoch": 0.5125815470643057,
+      "grad_norm": 0.4417482867433448,
+      "learning_rate": 4.606489471867449e-05,
+      "loss": 0.4293,
+      "step": 550
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 0.34875514025764925,
+      "learning_rate": 4.597859855022437e-05,
+      "loss": 0.4236,
+      "step": 555
+    },
+    {
+      "epoch": 0.5219012115563839,
+      "grad_norm": 0.4108051872320117,
+      "learning_rate": 4.589230238177425e-05,
+      "loss": 0.4445,
+      "step": 560
+    },
+    {
+      "epoch": 0.5265610438024231,
+      "grad_norm": 0.44712854316017836,
+      "learning_rate": 4.5806006213324134e-05,
+      "loss": 0.422,
+      "step": 565
+    },
+    {
+      "epoch": 0.5312208760484622,
+      "grad_norm": 0.5074461667448572,
+      "learning_rate": 4.5719710044874006e-05,
+      "loss": 0.4218,
+      "step": 570
+    },
+    {
+      "epoch": 0.5358807082945014,
+      "grad_norm": 0.40474852553762236,
+      "learning_rate": 4.563341387642389e-05,
+      "loss": 0.428,
+      "step": 575
+    },
+    {
+      "epoch": 0.5405405405405406,
+      "grad_norm": 0.35760028403552246,
+      "learning_rate": 4.554711770797377e-05,
+      "loss": 0.4281,
+      "step": 580
+    },
+    {
+      "epoch": 0.5452003727865797,
+      "grad_norm": 0.43397068927843285,
+      "learning_rate": 4.546082153952365e-05,
+      "loss": 0.4251,
+      "step": 585
+    },
+    {
+      "epoch": 0.5498602050326188,
+      "grad_norm": 0.39007262792966735,
+      "learning_rate": 4.5374525371073526e-05,
+      "loss": 0.4213,
+      "step": 590
+    },
+    {
+      "epoch": 0.554520037278658,
+      "grad_norm": 0.36154265093260674,
+      "learning_rate": 4.5288229202623404e-05,
+      "loss": 0.4236,
+      "step": 595
+    },
+    {
+      "epoch": 0.5591798695246971,
+      "grad_norm": 0.40445512036092346,
+      "learning_rate": 4.520193303417328e-05,
+      "loss": 0.4237,
+      "step": 600
+    },
+    {
+      "epoch": 0.5638397017707363,
+      "grad_norm": 0.3575679279572506,
+      "learning_rate": 4.511563686572317e-05,
+      "loss": 0.4226,
+      "step": 605
+    },
+    {
+      "epoch": 0.5684995340167754,
+      "grad_norm": 0.42859866287384585,
+      "learning_rate": 4.5029340697273045e-05,
+      "loss": 0.4337,
+      "step": 610
+    },
+    {
+      "epoch": 0.5731593662628145,
+      "grad_norm": 0.3398059428883966,
+      "learning_rate": 4.4943044528822923e-05,
+      "loss": 0.4202,
+      "step": 615
+    },
+    {
+      "epoch": 0.5778191985088537,
+      "grad_norm": 0.4329692306704453,
+      "learning_rate": 4.48567483603728e-05,
+      "loss": 0.4172,
+      "step": 620
+    },
+    {
+      "epoch": 0.5824790307548928,
+      "grad_norm": 0.3697726336264495,
+      "learning_rate": 4.477045219192268e-05,
+      "loss": 0.439,
+      "step": 625
+    },
+    {
+      "epoch": 0.587138863000932,
+      "grad_norm": 0.36623031936246386,
+      "learning_rate": 4.4684156023472565e-05,
+      "loss": 0.4295,
+      "step": 630
+    },
+    {
+      "epoch": 0.5917986952469712,
+      "grad_norm": 0.474742638876247,
+      "learning_rate": 4.4597859855022436e-05,
+      "loss": 0.4341,
+      "step": 635
+    },
+    {
+      "epoch": 0.5964585274930102,
+      "grad_norm": 0.48429407467593777,
+      "learning_rate": 4.4511563686572315e-05,
+      "loss": 0.4248,
+      "step": 640
+    },
+    {
+      "epoch": 0.6011183597390494,
+      "grad_norm": 0.42926923766107844,
+      "learning_rate": 4.44252675181222e-05,
+      "loss": 0.4246,
+      "step": 645
+    },
+    {
+      "epoch": 0.6057781919850885,
+      "grad_norm": 0.41759175987652447,
+      "learning_rate": 4.433897134967208e-05,
+      "loss": 0.4333,
+      "step": 650
+    },
+    {
+      "epoch": 0.6104380242311277,
+      "grad_norm": 0.40591710297089845,
+      "learning_rate": 4.4252675181221956e-05,
+      "loss": 0.4308,
+      "step": 655
+    },
+    {
+      "epoch": 0.6150978564771669,
+      "grad_norm": 0.34151741261761837,
+      "learning_rate": 4.4166379012771834e-05,
+      "loss": 0.4127,
+      "step": 660
+    },
+    {
+      "epoch": 0.6197576887232059,
+      "grad_norm": 0.4739000530318965,
+      "learning_rate": 4.408008284432171e-05,
+      "loss": 0.4134,
+      "step": 665
+    },
+    {
+      "epoch": 0.6244175209692451,
+      "grad_norm": 0.40594442187572644,
+      "learning_rate": 4.39937866758716e-05,
+      "loss": 0.4119,
+      "step": 670
+    },
+    {
+      "epoch": 0.6290773532152842,
+      "grad_norm": 0.38741906022032196,
+      "learning_rate": 4.3907490507421476e-05,
+      "loss": 0.4205,
+      "step": 675
+    },
+    {
+      "epoch": 0.6337371854613234,
+      "grad_norm": 0.40827486654087697,
+      "learning_rate": 4.382119433897135e-05,
+      "loss": 0.4235,
+      "step": 680
+    },
+    {
+      "epoch": 0.6383970177073626,
+      "grad_norm": 0.4472633880572486,
+      "learning_rate": 4.373489817052123e-05,
+      "loss": 0.4332,
+      "step": 685
+    },
+    {
+      "epoch": 0.6430568499534017,
+      "grad_norm": 0.4294679057605266,
+      "learning_rate": 4.364860200207111e-05,
+      "loss": 0.4153,
+      "step": 690
+    },
+    {
+      "epoch": 0.6477166821994408,
+      "grad_norm": 0.40635265146514193,
+      "learning_rate": 4.356230583362099e-05,
+      "loss": 0.4207,
+      "step": 695
+    },
+    {
+      "epoch": 0.65237651444548,
+      "grad_norm": 0.3677110047703502,
+      "learning_rate": 4.347600966517087e-05,
+      "loss": 0.429,
+      "step": 700
+    },
+    {
+      "epoch": 0.6570363466915191,
+      "grad_norm": 0.39912181862304363,
+      "learning_rate": 4.3389713496720745e-05,
+      "loss": 0.4225,
+      "step": 705
+    },
+    {
+      "epoch": 0.6616961789375583,
+      "grad_norm": 0.4706777419313193,
+      "learning_rate": 4.330341732827063e-05,
+      "loss": 0.4231,
+      "step": 710
+    },
+    {
+      "epoch": 0.6663560111835974,
+      "grad_norm": 0.3960054145843989,
+      "learning_rate": 4.321712115982051e-05,
+      "loss": 0.4115,
+      "step": 715
+    },
+    {
+      "epoch": 0.6710158434296365,
+      "grad_norm": 0.3796810889210068,
+      "learning_rate": 4.3130824991370387e-05,
+      "loss": 0.4219,
+      "step": 720
+    },
+    {
+      "epoch": 0.6756756756756757,
+      "grad_norm": 0.3231682922426093,
+      "learning_rate": 4.3044528822920265e-05,
+      "loss": 0.4207,
+      "step": 725
+    },
+    {
+      "epoch": 0.6803355079217148,
+      "grad_norm": 0.4338919098644665,
+      "learning_rate": 4.295823265447014e-05,
+      "loss": 0.4156,
+      "step": 730
+    },
+    {
+      "epoch": 0.684995340167754,
+      "grad_norm": 0.3769886966857475,
+      "learning_rate": 4.287193648602002e-05,
+      "loss": 0.4195,
+      "step": 735
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 0.3380840539678925,
+      "learning_rate": 4.27856403175699e-05,
+      "loss": 0.413,
+      "step": 740
+    },
+    {
+      "epoch": 0.6943150046598322,
+      "grad_norm": 0.36717339806709604,
+      "learning_rate": 4.269934414911978e-05,
+      "loss": 0.4089,
+      "step": 745
+    },
+    {
+      "epoch": 0.6989748369058714,
+      "grad_norm": 0.4168652698432254,
+      "learning_rate": 4.261304798066966e-05,
+      "loss": 0.4136,
+      "step": 750
+    },
+    {
+      "epoch": 0.7036346691519105,
+      "grad_norm": 0.3690477723388078,
+      "learning_rate": 4.252675181221954e-05,
+      "loss": 0.416,
+      "step": 755
+    },
+    {
+      "epoch": 0.7082945013979497,
+      "grad_norm": 0.4242455380044557,
+      "learning_rate": 4.244045564376942e-05,
+      "loss": 0.4191,
+      "step": 760
+    },
+    {
+      "epoch": 0.7129543336439889,
+      "grad_norm": 0.3735978801523571,
+      "learning_rate": 4.23541594753193e-05,
+      "loss": 0.4031,
+      "step": 765
+    },
+    {
+      "epoch": 0.7176141658900279,
+      "grad_norm": 0.3541170842431973,
+      "learning_rate": 4.2267863306869176e-05,
+      "loss": 0.4125,
+      "step": 770
+    },
+    {
+      "epoch": 0.7222739981360671,
+      "grad_norm": 0.3904517316938196,
+      "learning_rate": 4.2181567138419054e-05,
+      "loss": 0.425,
+      "step": 775
+    },
+    {
+      "epoch": 0.7269338303821062,
+      "grad_norm": 0.3553979878523653,
+      "learning_rate": 4.209527096996894e-05,
+      "loss": 0.4095,
+      "step": 780
+    },
+    {
+      "epoch": 0.7315936626281454,
+      "grad_norm": 0.3993958111614743,
+      "learning_rate": 4.200897480151881e-05,
+      "loss": 0.4168,
+      "step": 785
+    },
+    {
+      "epoch": 0.7362534948741846,
+      "grad_norm": 0.4854725144908592,
+      "learning_rate": 4.1922678633068695e-05,
+      "loss": 0.4201,
+      "step": 790
+    },
+    {
+      "epoch": 0.7409133271202236,
+      "grad_norm": 0.4093466282869362,
+      "learning_rate": 4.1836382464618573e-05,
+      "loss": 0.4144,
+      "step": 795
+    },
+    {
+      "epoch": 0.7455731593662628,
+      "grad_norm": 0.3802224792673468,
+      "learning_rate": 4.175008629616845e-05,
+      "loss": 0.4126,
+      "step": 800
+    },
+    {
+      "epoch": 0.750232991612302,
+      "grad_norm": 0.36566042296095735,
+      "learning_rate": 4.166379012771833e-05,
+      "loss": 0.4124,
+      "step": 805
+    },
+    {
+      "epoch": 0.7548928238583411,
+      "grad_norm": 0.33821479986680525,
+      "learning_rate": 4.157749395926821e-05,
+      "loss": 0.4307,
+      "step": 810
+    },
+    {
+      "epoch": 0.7595526561043803,
+      "grad_norm": 0.3717989938649144,
+      "learning_rate": 4.1491197790818086e-05,
+      "loss": 0.405,
+      "step": 815
+    },
+    {
+      "epoch": 0.7642124883504194,
+      "grad_norm": 0.5365115867656194,
+      "learning_rate": 4.140490162236797e-05,
+      "loss": 0.4203,
+      "step": 820
+    },
+    {
+      "epoch": 0.7688723205964585,
+      "grad_norm": 0.40471466095870456,
+      "learning_rate": 4.131860545391785e-05,
+      "loss": 0.4124,
+      "step": 825
+    },
+    {
+      "epoch": 0.7735321528424977,
+      "grad_norm": 0.38714529094413536,
+      "learning_rate": 4.123230928546773e-05,
+      "loss": 0.4073,
+      "step": 830
+    },
+    {
+      "epoch": 0.7781919850885368,
+      "grad_norm": 0.3949031763600824,
+      "learning_rate": 4.1146013117017606e-05,
+      "loss": 0.4143,
+      "step": 835
+    },
+    {
+      "epoch": 0.782851817334576,
+      "grad_norm": 0.4352023815662388,
+      "learning_rate": 4.1059716948567484e-05,
+      "loss": 0.4174,
+      "step": 840
+    },
+    {
+      "epoch": 0.7875116495806151,
+      "grad_norm": 0.4358406385453878,
+      "learning_rate": 4.097342078011737e-05,
+      "loss": 0.4254,
+      "step": 845
+    },
+    {
+      "epoch": 0.7921714818266542,
+      "grad_norm": 0.38366190649635135,
+      "learning_rate": 4.088712461166724e-05,
+      "loss": 0.4151,
+      "step": 850
+    },
+    {
+      "epoch": 0.7968313140726934,
+      "grad_norm": 0.5018848204235326,
+      "learning_rate": 4.080082844321712e-05,
+      "loss": 0.4171,
+      "step": 855
+    },
+    {
+      "epoch": 0.8014911463187325,
+      "grad_norm": 0.3758022403546901,
+      "learning_rate": 4.0714532274767004e-05,
+      "loss": 0.4241,
+      "step": 860
+    },
+    {
+      "epoch": 0.8061509785647717,
+      "grad_norm": 0.3562876865702164,
+      "learning_rate": 4.062823610631688e-05,
+      "loss": 0.405,
+      "step": 865
+    },
+    {
+      "epoch": 0.8108108108108109,
+      "grad_norm": 0.32964459523802225,
+      "learning_rate": 4.054193993786676e-05,
+      "loss": 0.4195,
+      "step": 870
+    },
+    {
+      "epoch": 0.8154706430568499,
+      "grad_norm": 0.3902217678069233,
+      "learning_rate": 4.045564376941664e-05,
+      "loss": 0.415,
+      "step": 875
+    },
+    {
+      "epoch": 0.8201304753028891,
+      "grad_norm": 0.6740352106087266,
+      "learning_rate": 4.036934760096652e-05,
+      "loss": 0.4065,
+      "step": 880
+    },
+    {
+      "epoch": 0.8247903075489282,
+      "grad_norm": 0.433837350976653,
+      "learning_rate": 4.02830514325164e-05,
+      "loss": 0.4206,
+      "step": 885
+    },
+    {
+      "epoch": 0.8294501397949674,
+      "grad_norm": 0.44183852945895624,
+      "learning_rate": 4.019675526406628e-05,
+      "loss": 0.4072,
+      "step": 890
+    },
+    {
+      "epoch": 0.8341099720410066,
+      "grad_norm": 0.34981947402981967,
+      "learning_rate": 4.011045909561615e-05,
+      "loss": 0.4267,
+      "step": 895
+    },
+    {
+      "epoch": 0.8387698042870456,
+      "grad_norm": 0.3704822813418764,
+      "learning_rate": 4.0024162927166037e-05,
+      "loss": 0.4093,
+      "step": 900
+    },
+    {
+      "epoch": 0.8434296365330848,
+      "grad_norm": 0.3440227445259439,
+      "learning_rate": 3.9937866758715915e-05,
+      "loss": 0.4109,
+      "step": 905
+    },
+    {
+      "epoch": 0.848089468779124,
+      "grad_norm": 0.46842513834829896,
+      "learning_rate": 3.98515705902658e-05,
+      "loss": 0.4169,
+      "step": 910
+    },
+    {
+      "epoch": 0.8527493010251631,
+      "grad_norm": 0.5632219425437093,
+      "learning_rate": 3.976527442181567e-05,
+      "loss": 0.4096,
+      "step": 915
+    },
+    {
+      "epoch": 0.8574091332712023,
+      "grad_norm": 0.4099216183303217,
+      "learning_rate": 3.967897825336555e-05,
+      "loss": 0.4162,
+      "step": 920
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 0.3475937062948154,
+      "learning_rate": 3.9592682084915434e-05,
+      "loss": 0.403,
+      "step": 925
+    },
+    {
+      "epoch": 0.8667287977632805,
+      "grad_norm": 0.4832320578936911,
+      "learning_rate": 3.950638591646531e-05,
+      "loss": 0.4025,
+      "step": 930
+    },
+    {
+      "epoch": 0.8713886300093197,
+      "grad_norm": 0.361390732321457,
+      "learning_rate": 3.942008974801519e-05,
+      "loss": 0.414,
+      "step": 935
+    },
+    {
+      "epoch": 0.8760484622553588,
+      "grad_norm": 0.2938396873991374,
+      "learning_rate": 3.933379357956507e-05,
+      "loss": 0.4096,
+      "step": 940
+    },
+    {
+      "epoch": 0.880708294501398,
+      "grad_norm": 0.4195411837550818,
+      "learning_rate": 3.924749741111495e-05,
+      "loss": 0.4119,
+      "step": 945
+    },
+    {
+      "epoch": 0.8853681267474371,
+      "grad_norm": 0.48658637807442856,
+      "learning_rate": 3.916120124266483e-05,
+      "loss": 0.4197,
+      "step": 950
+    },
+    {
+      "epoch": 0.8900279589934762,
+      "grad_norm": 0.4260894598328252,
+      "learning_rate": 3.9074905074214704e-05,
+      "loss": 0.4187,
+      "step": 955
+    },
+    {
+      "epoch": 0.8946877912395154,
+      "grad_norm": 0.34451778618434714,
+      "learning_rate": 3.898860890576458e-05,
+      "loss": 0.4138,
+      "step": 960
+    },
+    {
+      "epoch": 0.8993476234855545,
+      "grad_norm": 0.4365094696793227,
+      "learning_rate": 3.890231273731447e-05,
+      "loss": 0.4055,
+      "step": 965
+    },
+    {
+      "epoch": 0.9040074557315937,
+      "grad_norm": 0.3248156521748473,
+      "learning_rate": 3.8816016568864345e-05,
+      "loss": 0.4033,
+      "step": 970
+    },
+    {
+      "epoch": 0.9086672879776329,
+      "grad_norm": 0.32381986617162706,
+      "learning_rate": 3.8729720400414224e-05,
+      "loss": 0.4222,
+      "step": 975
+    },
+    {
+      "epoch": 0.9133271202236719,
+      "grad_norm": 0.400448857756024,
+      "learning_rate": 3.86434242319641e-05,
+      "loss": 0.4093,
+      "step": 980
+    },
+    {
+      "epoch": 0.9179869524697111,
+      "grad_norm": 0.5091133203813641,
+      "learning_rate": 3.855712806351398e-05,
+      "loss": 0.4124,
+      "step": 985
+    },
+    {
+      "epoch": 0.9226467847157502,
+      "grad_norm": 0.43202295251303324,
+      "learning_rate": 3.8470831895063865e-05,
+      "loss": 0.4088,
+      "step": 990
+    },
+    {
+      "epoch": 0.9273066169617894,
+      "grad_norm": 0.364710789588561,
+      "learning_rate": 3.838453572661374e-05,
+      "loss": 0.4046,
+      "step": 995
+    },
+    {
+      "epoch": 0.9319664492078286,
+      "grad_norm": 0.4716388737566944,
+      "learning_rate": 3.8298239558163615e-05,
+      "loss": 0.4075,
+      "step": 1000
+    },
+    {
+      "epoch": 0.9366262814538676,
+      "grad_norm": 0.6143023930914768,
+      "learning_rate": 3.82119433897135e-05,
+      "loss": 0.4151,
+      "step": 1005
+    },
+    {
+      "epoch": 0.9412861136999068,
+      "grad_norm": 0.4045209641384314,
+      "learning_rate": 3.812564722126338e-05,
+      "loss": 0.4061,
+      "step": 1010
+    },
+    {
+      "epoch": 0.9459459459459459,
+      "grad_norm": 0.4374289349183326,
+      "learning_rate": 3.8039351052813256e-05,
+      "loss": 0.4149,
+      "step": 1015
+    },
+    {
+      "epoch": 0.9506057781919851,
+      "grad_norm": 0.4186126910891237,
+      "learning_rate": 3.7953054884363134e-05,
+      "loss": 0.4082,
+      "step": 1020
+    },
+    {
+      "epoch": 0.9552656104380243,
+      "grad_norm": 0.3724784755826059,
+      "learning_rate": 3.786675871591301e-05,
+      "loss": 0.4116,
+      "step": 1025
+    },
+    {
+      "epoch": 0.9599254426840633,
+      "grad_norm": 0.3162214191511966,
+      "learning_rate": 3.77804625474629e-05,
+      "loss": 0.4165,
+      "step": 1030
+    },
+    {
+      "epoch": 0.9645852749301025,
+      "grad_norm": 0.44560720068567894,
+      "learning_rate": 3.7694166379012776e-05,
+      "loss": 0.4086,
+      "step": 1035
+    },
+    {
+      "epoch": 0.9692451071761417,
+      "grad_norm": 0.3715493465051349,
+      "learning_rate": 3.7607870210562654e-05,
+      "loss": 0.406,
+      "step": 1040
+    },
+    {
+      "epoch": 0.9739049394221808,
+      "grad_norm": 0.2864953938663117,
+      "learning_rate": 3.752157404211253e-05,
+      "loss": 0.3984,
+      "step": 1045
+    },
+    {
+      "epoch": 0.97856477166822,
+      "grad_norm": 0.4430521911068819,
+      "learning_rate": 3.743527787366241e-05,
+      "loss": 0.428,
+      "step": 1050
+    },
+    {
+      "epoch": 0.983224603914259,
+      "grad_norm": 0.39380676782546514,
+      "learning_rate": 3.734898170521229e-05,
+      "loss": 0.4058,
+      "step": 1055
+    },
+    {
+      "epoch": 0.9878844361602982,
+      "grad_norm": 0.45421105453517185,
+      "learning_rate": 3.7262685536762174e-05,
+      "loss": 0.4048,
+      "step": 1060
+    },
+    {
+      "epoch": 0.9925442684063374,
+      "grad_norm": 0.3188847720393995,
+      "learning_rate": 3.7176389368312045e-05,
+      "loss": 0.4088,
+      "step": 1065
+    },
+    {
+      "epoch": 0.9972041006523765,
+      "grad_norm": 0.3979525304245011,
+      "learning_rate": 3.709009319986193e-05,
+      "loss": 0.409,
+      "step": 1070
+    },
+    {
+      "epoch": 1.0018639328984156,
+      "grad_norm": 0.364266540801231,
+      "learning_rate": 3.700379703141181e-05,
+      "loss": 0.3739,
+      "step": 1075
+    },
+    {
+      "epoch": 1.0065237651444547,
+      "grad_norm": 0.30445308416956596,
+      "learning_rate": 3.6917500862961687e-05,
+      "loss": 0.3498,
+      "step": 1080
+    },
+    {
+      "epoch": 1.011183597390494,
+      "grad_norm": 0.4080914841436169,
+      "learning_rate": 3.6831204694511565e-05,
+      "loss": 0.3564,
+      "step": 1085
+    },
+    {
+      "epoch": 1.015843429636533,
+      "grad_norm": 0.3649554671829125,
+      "learning_rate": 3.674490852606144e-05,
+      "loss": 0.3637,
+      "step": 1090
+    },
+    {
+      "epoch": 1.0205032618825722,
+      "grad_norm": 0.3163559100974115,
+      "learning_rate": 3.665861235761132e-05,
+      "loss": 0.349,
+      "step": 1095
+    },
+    {
+      "epoch": 1.0251630941286114,
+      "grad_norm": 0.36263837002354715,
+      "learning_rate": 3.6572316189161206e-05,
+      "loss": 0.3403,
+      "step": 1100
+    },
+    {
+      "epoch": 1.0298229263746506,
+      "grad_norm": 0.3092678358686789,
+      "learning_rate": 3.6486020020711085e-05,
+      "loss": 0.3465,
+      "step": 1105
+    },
+    {
+      "epoch": 1.0344827586206897,
+      "grad_norm": 0.3141892800399063,
+      "learning_rate": 3.639972385226096e-05,
+      "loss": 0.347,
+      "step": 1110
+    },
+    {
+      "epoch": 1.0391425908667289,
+      "grad_norm": 0.31295163640755363,
+      "learning_rate": 3.631342768381084e-05,
+      "loss": 0.3471,
+      "step": 1115
+    },
+    {
+      "epoch": 1.0438024231127678,
+      "grad_norm": 0.3401750953887306,
+      "learning_rate": 3.622713151536072e-05,
+      "loss": 0.3473,
+      "step": 1120
+    },
+    {
+      "epoch": 1.048462255358807,
+      "grad_norm": 0.31195615894221534,
+      "learning_rate": 3.6140835346910604e-05,
+      "loss": 0.3507,
+      "step": 1125
+    },
+    {
+      "epoch": 1.0531220876048462,
+      "grad_norm": 0.32979807481995643,
+      "learning_rate": 3.6054539178460476e-05,
+      "loss": 0.347,
+      "step": 1130
+    },
+    {
+      "epoch": 1.0577819198508853,
+      "grad_norm": 0.3759497094197671,
+      "learning_rate": 3.5968243010010354e-05,
+      "loss": 0.3635,
+      "step": 1135
+    },
+    {
+      "epoch": 1.0624417520969245,
+      "grad_norm": 0.3276972337838372,
+      "learning_rate": 3.588194684156024e-05,
+      "loss": 0.3546,
+      "step": 1140
+    },
+    {
+      "epoch": 1.0671015843429636,
+      "grad_norm": 0.3248203280892395,
+      "learning_rate": 3.579565067311012e-05,
+      "loss": 0.3461,
+      "step": 1145
+    },
+    {
+      "epoch": 1.0717614165890028,
+      "grad_norm": 0.3049982165023373,
+      "learning_rate": 3.5709354504659995e-05,
+      "loss": 0.3425,
+      "step": 1150
+    },
+    {
+      "epoch": 1.076421248835042,
+      "grad_norm": 0.3359497674452099,
+      "learning_rate": 3.5623058336209874e-05,
+      "loss": 0.3535,
+      "step": 1155
+    },
+    {
+      "epoch": 1.0810810810810811,
+      "grad_norm": 0.3158582105691738,
+      "learning_rate": 3.553676216775975e-05,
+      "loss": 0.3448,
+      "step": 1160
+    },
+    {
+      "epoch": 1.0857409133271203,
+      "grad_norm": 0.29574757335640145,
+      "learning_rate": 3.545046599930964e-05,
+      "loss": 0.3565,
+      "step": 1165
+    },
+    {
+      "epoch": 1.0904007455731595,
+      "grad_norm": 0.31108264753865356,
+      "learning_rate": 3.536416983085951e-05,
+      "loss": 0.3583,
+      "step": 1170
+    },
+    {
+      "epoch": 1.0950605778191984,
+      "grad_norm": 0.36037209266568915,
+      "learning_rate": 3.5277873662409386e-05,
+      "loss": 0.3448,
+      "step": 1175
+    },
+    {
+      "epoch": 1.0997204100652376,
+      "grad_norm": 0.3618383231487912,
+      "learning_rate": 3.519157749395927e-05,
+      "loss": 0.3507,
+      "step": 1180
+    },
+    {
+      "epoch": 1.1043802423112767,
+      "grad_norm": 0.4175780783750628,
+      "learning_rate": 3.510528132550915e-05,
+      "loss": 0.3479,
+      "step": 1185
+    },
+    {
+      "epoch": 1.109040074557316,
+      "grad_norm": 0.3599957546238691,
+      "learning_rate": 3.501898515705903e-05,
+      "loss": 0.3503,
+      "step": 1190
+    },
+    {
+      "epoch": 1.113699906803355,
+      "grad_norm": 0.3501599765642264,
+      "learning_rate": 3.4932688988608906e-05,
+      "loss": 0.3612,
+      "step": 1195
+    },
+    {
+      "epoch": 1.1183597390493942,
+      "grad_norm": 0.5599389993107197,
+      "learning_rate": 3.4846392820158784e-05,
+      "loss": 0.3512,
+      "step": 1200
+    },
+    {
+      "epoch": 1.1230195712954334,
+      "grad_norm": 0.39235842650287267,
+      "learning_rate": 3.476009665170867e-05,
+      "loss": 0.3503,
+      "step": 1205
+    },
+    {
+      "epoch": 1.1276794035414726,
+      "grad_norm": 0.35965537290287625,
+      "learning_rate": 3.467380048325855e-05,
+      "loss": 0.3505,
+      "step": 1210
+    },
+    {
+      "epoch": 1.1323392357875117,
+      "grad_norm": 0.3147552390929704,
+      "learning_rate": 3.458750431480842e-05,
+      "loss": 0.3556,
+      "step": 1215
+    },
+    {
+      "epoch": 1.1369990680335509,
+      "grad_norm": 0.35667525578447856,
+      "learning_rate": 3.4501208146358304e-05,
+      "loss": 0.3551,
+      "step": 1220
+    },
+    {
+      "epoch": 1.14165890027959,
+      "grad_norm": 0.3143522174903333,
+      "learning_rate": 3.441491197790818e-05,
+      "loss": 0.3506,
+      "step": 1225
+    },
+    {
+      "epoch": 1.146318732525629,
+      "grad_norm": 0.2953129277848881,
+      "learning_rate": 3.432861580945806e-05,
+      "loss": 0.3519,
+      "step": 1230
+    },
+    {
+      "epoch": 1.1509785647716682,
+      "grad_norm": 0.3201828959982261,
+      "learning_rate": 3.424231964100794e-05,
+      "loss": 0.351,
+      "step": 1235
+    },
+    {
+      "epoch": 1.1556383970177073,
+      "grad_norm": 0.35805902539565565,
+      "learning_rate": 3.415602347255782e-05,
+      "loss": 0.3496,
+      "step": 1240
+    },
+    {
+      "epoch": 1.1602982292637465,
+      "grad_norm": 0.2844778506189992,
+      "learning_rate": 3.40697273041077e-05,
+      "loss": 0.3567,
+      "step": 1245
+    },
+    {
+      "epoch": 1.1649580615097856,
+      "grad_norm": 0.3932646163355652,
+      "learning_rate": 3.398343113565758e-05,
+      "loss": 0.3508,
+      "step": 1250
+    },
+    {
+      "epoch": 1.1696178937558248,
+      "grad_norm": 0.27258390887519407,
+      "learning_rate": 3.389713496720746e-05,
+      "loss": 0.3296,
+      "step": 1255
+    },
+    {
+      "epoch": 1.174277726001864,
+      "grad_norm": 0.2775303977010617,
+      "learning_rate": 3.381083879875734e-05,
+      "loss": 0.3491,
+      "step": 1260
+    },
+    {
+      "epoch": 1.1789375582479031,
+      "grad_norm": 0.3028564721433078,
+      "learning_rate": 3.3724542630307215e-05,
+      "loss": 0.3456,
+      "step": 1265
+    },
+    {
+      "epoch": 1.1835973904939423,
+      "grad_norm": 0.32923905528364905,
+      "learning_rate": 3.363824646185709e-05,
+      "loss": 0.3555,
+      "step": 1270
+    },
+    {
+      "epoch": 1.1882572227399812,
+      "grad_norm": 0.2760919775973278,
+      "learning_rate": 3.355195029340698e-05,
+      "loss": 0.3489,
+      "step": 1275
+    },
+    {
+      "epoch": 1.1929170549860204,
+      "grad_norm": 0.3155488629032025,
+      "learning_rate": 3.346565412495685e-05,
+      "loss": 0.3597,
+      "step": 1280
+    },
+    {
+      "epoch": 1.1975768872320596,
+      "grad_norm": 0.34778125198320686,
+      "learning_rate": 3.3379357956506735e-05,
+      "loss": 0.3605,
+      "step": 1285
+    },
+    {
+      "epoch": 1.2022367194780987,
+      "grad_norm": 0.3591989412785019,
+      "learning_rate": 3.329306178805661e-05,
+      "loss": 0.3439,
+      "step": 1290
+    },
+    {
+      "epoch": 1.206896551724138,
+      "grad_norm": 0.338584951279966,
+      "learning_rate": 3.320676561960649e-05,
+      "loss": 0.3486,
+      "step": 1295
+    },
+    {
+      "epoch": 1.211556383970177,
+      "grad_norm": 0.44502827017418983,
+      "learning_rate": 3.312046945115637e-05,
+      "loss": 0.3574,
+      "step": 1300
+    },
+    {
+      "epoch": 1.2162162162162162,
+      "grad_norm": 0.37848494564678853,
+      "learning_rate": 3.303417328270625e-05,
+      "loss": 0.3554,
+      "step": 1305
+    },
+    {
+      "epoch": 1.2208760484622554,
+      "grad_norm": 0.2991926203877817,
+      "learning_rate": 3.2947877114256126e-05,
+      "loss": 0.3448,
+      "step": 1310
+    },
+    {
+      "epoch": 1.2255358807082946,
+      "grad_norm": 0.33851023261699276,
+      "learning_rate": 3.286158094580601e-05,
+      "loss": 0.3648,
+      "step": 1315
+    },
+    {
+      "epoch": 1.2301957129543337,
+      "grad_norm": 0.27617927948407645,
+      "learning_rate": 3.277528477735589e-05,
+      "loss": 0.3505,
+      "step": 1320
+    },
+    {
+      "epoch": 1.2348555452003729,
+      "grad_norm": 0.31943221432421715,
+      "learning_rate": 3.268898860890577e-05,
+      "loss": 0.3575,
+      "step": 1325
+    },
+    {
+      "epoch": 1.2395153774464118,
+      "grad_norm": 0.36090830340862545,
+      "learning_rate": 3.2602692440455645e-05,
+      "loss": 0.3459,
+      "step": 1330
+    },
+    {
+      "epoch": 1.244175209692451,
+      "grad_norm": 0.2869837155345359,
+      "learning_rate": 3.2516396272005524e-05,
+      "loss": 0.3483,
+      "step": 1335
+    },
+    {
+      "epoch": 1.2488350419384902,
+      "grad_norm": 0.4574789609384581,
+      "learning_rate": 3.243010010355541e-05,
+      "loss": 0.3537,
+      "step": 1340
+    },
+    {
+      "epoch": 1.2534948741845293,
+      "grad_norm": 0.34435517087930184,
+      "learning_rate": 3.234380393510528e-05,
+      "loss": 0.3412,
+      "step": 1345
+    },
+    {
+      "epoch": 1.2581547064305685,
+      "grad_norm": 0.28787047833310464,
+      "learning_rate": 3.225750776665516e-05,
+      "loss": 0.3481,
+      "step": 1350
+    },
+    {
+      "epoch": 1.2628145386766076,
+      "grad_norm": 0.28433333546839407,
+      "learning_rate": 3.217121159820504e-05,
+      "loss": 0.3581,
+      "step": 1355
+    },
+    {
+      "epoch": 1.2674743709226468,
+      "grad_norm": 0.2837264128964714,
+      "learning_rate": 3.208491542975492e-05,
+      "loss": 0.3475,
+      "step": 1360
+    },
+    {
+      "epoch": 1.272134203168686,
+      "grad_norm": 0.2783632546885593,
+      "learning_rate": 3.19986192613048e-05,
+      "loss": 0.3528,
+      "step": 1365
+    },
+    {
+      "epoch": 1.2767940354147251,
+      "grad_norm": 0.32499690495431244,
+      "learning_rate": 3.191232309285468e-05,
+      "loss": 0.352,
+      "step": 1370
+    },
+    {
+      "epoch": 1.281453867660764,
+      "grad_norm": 0.3177042804556951,
+      "learning_rate": 3.1826026924404556e-05,
+      "loss": 0.3416,
+      "step": 1375
+    },
+    {
+      "epoch": 1.2861136999068035,
+      "grad_norm": 0.3082833632499775,
+      "learning_rate": 3.173973075595444e-05,
+      "loss": 0.3561,
+      "step": 1380
+    },
+    {
+      "epoch": 1.2907735321528424,
+      "grad_norm": 0.3192882141973088,
+      "learning_rate": 3.165343458750431e-05,
+      "loss": 0.358,
+      "step": 1385
+    },
+    {
+      "epoch": 1.2954333643988816,
+      "grad_norm": 0.35083440823453566,
+      "learning_rate": 3.156713841905419e-05,
+      "loss": 0.3482,
+      "step": 1390
+    },
+    {
+      "epoch": 1.3000931966449207,
+      "grad_norm": 0.3028496199107332,
+      "learning_rate": 3.1480842250604076e-05,
+      "loss": 0.3524,
+      "step": 1395
+    },
+    {
+      "epoch": 1.30475302889096,
+      "grad_norm": 0.40238805661277516,
+      "learning_rate": 3.1394546082153954e-05,
+      "loss": 0.3465,
+      "step": 1400
+    },
+    {
+      "epoch": 1.309412861136999,
+      "grad_norm": 0.38264937163405993,
+      "learning_rate": 3.130824991370383e-05,
+      "loss": 0.3572,
+      "step": 1405
+    },
+    {
+      "epoch": 1.3140726933830382,
+      "grad_norm": 0.3476478264061866,
+      "learning_rate": 3.122195374525371e-05,
+      "loss": 0.3452,
+      "step": 1410
+    },
+    {
+      "epoch": 1.3187325256290774,
+      "grad_norm": 0.29460741785810235,
+      "learning_rate": 3.113565757680359e-05,
+      "loss": 0.3528,
+      "step": 1415
+    },
+    {
+      "epoch": 1.3233923578751166,
+      "grad_norm": 0.3105487189575632,
+      "learning_rate": 3.1049361408353474e-05,
+      "loss": 0.3403,
+      "step": 1420
+    },
+    {
+      "epoch": 1.3280521901211557,
+      "grad_norm": 0.24970556065463576,
+      "learning_rate": 3.096306523990335e-05,
+      "loss": 0.3481,
+      "step": 1425
+    },
+    {
+      "epoch": 1.3327120223671947,
+      "grad_norm": 0.26478763592488186,
+      "learning_rate": 3.0876769071453223e-05,
+      "loss": 0.3567,
+      "step": 1430
+    },
+    {
+      "epoch": 1.337371854613234,
+      "grad_norm": 0.3166787826392816,
+      "learning_rate": 3.079047290300311e-05,
+      "loss": 0.3549,
+      "step": 1435
+    },
+    {
+      "epoch": 1.342031686859273,
+      "grad_norm": 0.3370862593924141,
+      "learning_rate": 3.070417673455299e-05,
+      "loss": 0.3532,
+      "step": 1440
+    },
+    {
+      "epoch": 1.3466915191053122,
+      "grad_norm": 0.28943779202148834,
+      "learning_rate": 3.061788056610287e-05,
+      "loss": 0.3578,
+      "step": 1445
+    },
+    {
+      "epoch": 1.3513513513513513,
+      "grad_norm": 0.3305096681079534,
+      "learning_rate": 3.053158439765274e-05,
+      "loss": 0.3434,
+      "step": 1450
+    },
+    {
+      "epoch": 1.3560111835973905,
+      "grad_norm": 0.30770425179734145,
+      "learning_rate": 3.0445288229202625e-05,
+      "loss": 0.3513,
+      "step": 1455
+    },
+    {
+      "epoch": 1.3606710158434296,
+      "grad_norm": 0.29217231918736536,
+      "learning_rate": 3.0358992060752506e-05,
+      "loss": 0.357,
+      "step": 1460
+    },
+    {
+      "epoch": 1.3653308480894688,
+      "grad_norm": 0.30497087668658185,
+      "learning_rate": 3.027269589230238e-05,
+      "loss": 0.3512,
+      "step": 1465
+    },
+    {
+      "epoch": 1.369990680335508,
+      "grad_norm": 0.3169652144095064,
+      "learning_rate": 3.018639972385226e-05,
+      "loss": 0.3485,
+      "step": 1470
+    },
+    {
+      "epoch": 1.3746505125815471,
+      "grad_norm": 0.33878452557048316,
+      "learning_rate": 3.010010355540214e-05,
+      "loss": 0.3501,
+      "step": 1475
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 0.25128935811284947,
+      "learning_rate": 3.001380738695202e-05,
+      "loss": 0.3342,
+      "step": 1480
+    },
+    {
+      "epoch": 1.3839701770736252,
+      "grad_norm": 0.2807928725658594,
+      "learning_rate": 2.99275112185019e-05,
+      "loss": 0.3515,
+      "step": 1485
+    },
+    {
+      "epoch": 1.3886300093196646,
+      "grad_norm": 0.32627986332317066,
+      "learning_rate": 2.984121505005178e-05,
+      "loss": 0.3564,
+      "step": 1490
+    },
+    {
+      "epoch": 1.3932898415657036,
+      "grad_norm": 0.34700278304006454,
+      "learning_rate": 2.9754918881601657e-05,
+      "loss": 0.3647,
+      "step": 1495
+    },
+    {
+      "epoch": 1.3979496738117427,
+      "grad_norm": 0.33291240407016004,
+      "learning_rate": 2.966862271315154e-05,
+      "loss": 0.3608,
+      "step": 1500
+    },
+    {
+      "epoch": 1.402609506057782,
+      "grad_norm": 0.36233550445620966,
+      "learning_rate": 2.9582326544701417e-05,
+      "loss": 0.3576,
+      "step": 1505
+    },
+    {
+      "epoch": 1.407269338303821,
+      "grad_norm": 0.299037421676663,
+      "learning_rate": 2.9496030376251292e-05,
+      "loss": 0.35,
+      "step": 1510
+    },
+    {
+      "epoch": 1.4119291705498602,
+      "grad_norm": 0.2779558872885985,
+      "learning_rate": 2.9409734207801177e-05,
+      "loss": 0.3536,
+      "step": 1515
+    },
+    {
+      "epoch": 1.4165890027958994,
+      "grad_norm": 0.300917823810139,
+      "learning_rate": 2.9323438039351052e-05,
+      "loss": 0.3527,
+      "step": 1520
+    },
+    {
+      "epoch": 1.4212488350419386,
+      "grad_norm": 0.2914307296082574,
+      "learning_rate": 2.9237141870900937e-05,
+      "loss": 0.3487,
+      "step": 1525
+    },
+    {
+      "epoch": 1.4259086672879777,
+      "grad_norm": 0.2856071431198464,
+      "learning_rate": 2.9150845702450812e-05,
+      "loss": 0.3476,
+      "step": 1530
+    },
+    {
+      "epoch": 1.4305684995340169,
+      "grad_norm": 0.3129750830236417,
+      "learning_rate": 2.906454953400069e-05,
+      "loss": 0.3524,
+      "step": 1535
+    },
+    {
+      "epoch": 1.4352283317800558,
+      "grad_norm": 0.3179106281194418,
+      "learning_rate": 2.897825336555057e-05,
+      "loss": 0.3485,
+      "step": 1540
+    },
+    {
+      "epoch": 1.439888164026095,
+      "grad_norm": 0.38784207287547934,
+      "learning_rate": 2.889195719710045e-05,
+      "loss": 0.349,
+      "step": 1545
+    },
+    {
+      "epoch": 1.4445479962721341,
+      "grad_norm": 0.3753888528298807,
+      "learning_rate": 2.8805661028650328e-05,
+      "loss": 0.3451,
+      "step": 1550
+    },
+    {
+      "epoch": 1.4492078285181733,
+      "grad_norm": 0.41321946201235715,
+      "learning_rate": 2.871936486020021e-05,
+      "loss": 0.3426,
+      "step": 1555
+    },
+    {
+      "epoch": 1.4538676607642125,
+      "grad_norm": 0.3200312045653147,
+      "learning_rate": 2.8633068691750088e-05,
+      "loss": 0.3453,
+      "step": 1560
+    },
+    {
+      "epoch": 1.4585274930102516,
+      "grad_norm": 0.3307103194113536,
+      "learning_rate": 2.854677252329997e-05,
+      "loss": 0.3476,
+      "step": 1565
+    },
+    {
+      "epoch": 1.4631873252562908,
+      "grad_norm": 0.30755068927447626,
+      "learning_rate": 2.8460476354849848e-05,
+      "loss": 0.3546,
+      "step": 1570
+    },
+    {
+      "epoch": 1.46784715750233,
+      "grad_norm": 0.3595240127776032,
+      "learning_rate": 2.8374180186399723e-05,
+      "loss": 0.3546,
+      "step": 1575
+    },
+    {
+      "epoch": 1.4725069897483691,
+      "grad_norm": 0.3689101626664539,
+      "learning_rate": 2.8287884017949608e-05,
+      "loss": 0.3557,
+      "step": 1580
+    },
+    {
+      "epoch": 1.477166821994408,
+      "grad_norm": 0.3583075317481452,
+      "learning_rate": 2.8201587849499482e-05,
+      "loss": 0.3487,
+      "step": 1585
+    },
+    {
+      "epoch": 1.4818266542404475,
+      "grad_norm": 0.35602642963014874,
+      "learning_rate": 2.811529168104936e-05,
+      "loss": 0.3591,
+      "step": 1590
+    },
+    {
+      "epoch": 1.4864864864864864,
+      "grad_norm": 0.315787516045647,
+      "learning_rate": 2.8028995512599242e-05,
+      "loss": 0.3499,
+      "step": 1595
+    },
+    {
+      "epoch": 1.4911463187325256,
+      "grad_norm": 0.3590015577504981,
+      "learning_rate": 2.794269934414912e-05,
+      "loss": 0.3512,
+      "step": 1600
+    },
+    {
+      "epoch": 1.4958061509785647,
+      "grad_norm": 0.3122273467939273,
+      "learning_rate": 2.7856403175699002e-05,
+      "loss": 0.3557,
+      "step": 1605
+    },
+    {
+      "epoch": 1.500465983224604,
+      "grad_norm": 0.3542977601514441,
+      "learning_rate": 2.777010700724888e-05,
+      "loss": 0.3508,
+      "step": 1610
+    },
+    {
+      "epoch": 1.505125815470643,
+      "grad_norm": 0.3391938107105768,
+      "learning_rate": 2.768381083879876e-05,
+      "loss": 0.3458,
+      "step": 1615
+    },
+    {
+      "epoch": 1.5097856477166822,
+      "grad_norm": 0.2960461404701141,
+      "learning_rate": 2.759751467034864e-05,
+      "loss": 0.351,
+      "step": 1620
+    },
+    {
+      "epoch": 1.5144454799627214,
+      "grad_norm": 0.27438908752786395,
+      "learning_rate": 2.751121850189852e-05,
+      "loss": 0.3428,
+      "step": 1625
+    },
+    {
+      "epoch": 1.5191053122087603,
+      "grad_norm": 0.28849482254982545,
+      "learning_rate": 2.7424922333448393e-05,
+      "loss": 0.3366,
+      "step": 1630
+    },
+    {
+      "epoch": 1.5237651444547997,
+      "grad_norm": 0.3049772048163229,
+      "learning_rate": 2.7338626164998278e-05,
+      "loss": 0.3433,
+      "step": 1635
+    },
+    {
+      "epoch": 1.5284249767008387,
+      "grad_norm": 0.3325604613028116,
+      "learning_rate": 2.7252329996548153e-05,
+      "loss": 0.3463,
+      "step": 1640
+    },
+    {
+      "epoch": 1.533084808946878,
+      "grad_norm": 0.27987089540936305,
+      "learning_rate": 2.7166033828098038e-05,
+      "loss": 0.3614,
+      "step": 1645
+    },
+    {
+      "epoch": 1.537744641192917,
+      "grad_norm": 0.3045635963851511,
+      "learning_rate": 2.7079737659647913e-05,
+      "loss": 0.3505,
+      "step": 1650
+    },
+    {
+      "epoch": 1.5424044734389561,
+      "grad_norm": 0.3290983977476954,
+      "learning_rate": 2.699344149119779e-05,
+      "loss": 0.3572,
+      "step": 1655
+    },
+    {
+      "epoch": 1.5470643056849953,
+      "grad_norm": 0.28498739942206075,
+      "learning_rate": 2.6907145322747673e-05,
+      "loss": 0.3517,
+      "step": 1660
+    },
+    {
+      "epoch": 1.5517241379310345,
+      "grad_norm": 0.329718281320631,
+      "learning_rate": 2.682084915429755e-05,
+      "loss": 0.3425,
+      "step": 1665
+    },
+    {
+      "epoch": 1.5563839701770736,
+      "grad_norm": 0.3101714326036767,
+      "learning_rate": 2.673455298584743e-05,
+      "loss": 0.3473,
+      "step": 1670
+    },
+    {
+      "epoch": 1.5610438024231128,
+      "grad_norm": 0.31191112180013836,
+      "learning_rate": 2.664825681739731e-05,
+      "loss": 0.3549,
+      "step": 1675
+    },
+    {
+      "epoch": 1.565703634669152,
+      "grad_norm": 0.2903071511782437,
+      "learning_rate": 2.6561960648947186e-05,
+      "loss": 0.3465,
+      "step": 1680
+    },
+    {
+      "epoch": 1.570363466915191,
+      "grad_norm": 0.28707119756169747,
+      "learning_rate": 2.647566448049707e-05,
+      "loss": 0.3425,
+      "step": 1685
+    },
+    {
+      "epoch": 1.5750232991612303,
+      "grad_norm": 0.27354054458513866,
+      "learning_rate": 2.6389368312046945e-05,
+      "loss": 0.3547,
+      "step": 1690
+    },
+    {
+      "epoch": 1.5796831314072692,
+      "grad_norm": 0.3125368993905314,
+      "learning_rate": 2.6303072143596824e-05,
+      "loss": 0.3587,
+      "step": 1695
+    },
+    {
+      "epoch": 1.5843429636533086,
+      "grad_norm": 0.36883984027257033,
+      "learning_rate": 2.6216775975146705e-05,
+      "loss": 0.3511,
+      "step": 1700
+    },
+    {
+      "epoch": 1.5890027958993476,
+      "grad_norm": 0.28692323766609046,
+      "learning_rate": 2.6130479806696584e-05,
+      "loss": 0.3485,
+      "step": 1705
+    },
+    {
+      "epoch": 1.5936626281453867,
+      "grad_norm": 0.3481893519300557,
+      "learning_rate": 2.6044183638246462e-05,
+      "loss": 0.3541,
+      "step": 1710
+    },
+    {
+      "epoch": 1.598322460391426,
+      "grad_norm": 0.23239082555167248,
+      "learning_rate": 2.5957887469796343e-05,
+      "loss": 0.3434,
+      "step": 1715
+    },
+    {
+      "epoch": 1.602982292637465,
+      "grad_norm": 0.2999944857695801,
+      "learning_rate": 2.587159130134622e-05,
+      "loss": 0.348,
+      "step": 1720
+    },
+    {
+      "epoch": 1.6076421248835042,
+      "grad_norm": 0.2722028105267068,
+      "learning_rate": 2.5785295132896096e-05,
+      "loss": 0.3478,
+      "step": 1725
+    },
+    {
+      "epoch": 1.6123019571295434,
+      "grad_norm": 0.3090404281420746,
+      "learning_rate": 2.569899896444598e-05,
+      "loss": 0.3417,
+      "step": 1730
+    },
+    {
+      "epoch": 1.6169617893755825,
+      "grad_norm": 0.34151837735928975,
+      "learning_rate": 2.5612702795995856e-05,
+      "loss": 0.3535,
+      "step": 1735
+    },
+    {
+      "epoch": 1.6216216216216215,
+      "grad_norm": 0.3138996767589973,
+      "learning_rate": 2.552640662754574e-05,
+      "loss": 0.3471,
+      "step": 1740
+    },
+    {
+      "epoch": 1.6262814538676609,
+      "grad_norm": 0.2434520617717452,
+      "learning_rate": 2.5440110459095616e-05,
+      "loss": 0.3543,
+      "step": 1745
+    },
+    {
+      "epoch": 1.6309412861136998,
+      "grad_norm": 0.3320654632100791,
+      "learning_rate": 2.5353814290645494e-05,
+      "loss": 0.3527,
+      "step": 1750
+    },
+    {
+      "epoch": 1.6356011183597392,
+      "grad_norm": 0.30123970869727656,
+      "learning_rate": 2.5267518122195376e-05,
+      "loss": 0.3446,
+      "step": 1755
+    },
+    {
+      "epoch": 1.6402609506057781,
+      "grad_norm": 0.37186285485228093,
+      "learning_rate": 2.5181221953745254e-05,
+      "loss": 0.353,
+      "step": 1760
+    },
+    {
+      "epoch": 1.6449207828518173,
+      "grad_norm": 0.3206055010260991,
+      "learning_rate": 2.5094925785295132e-05,
+      "loss": 0.3544,
+      "step": 1765
+    },
+    {
+      "epoch": 1.6495806150978565,
+      "grad_norm": 0.2611046380695705,
+      "learning_rate": 2.5008629616845014e-05,
+      "loss": 0.3477,
+      "step": 1770
+    },
+    {
+      "epoch": 1.6542404473438956,
+      "grad_norm": 0.33682031679412877,
+      "learning_rate": 2.4922333448394892e-05,
+      "loss": 0.3403,
+      "step": 1775
+    },
+    {
+      "epoch": 1.6589002795899348,
+      "grad_norm": 0.38536041112111596,
+      "learning_rate": 2.483603727994477e-05,
+      "loss": 0.3503,
+      "step": 1780
+    },
+    {
+      "epoch": 1.6635601118359737,
+      "grad_norm": 0.34997202362889474,
+      "learning_rate": 2.4749741111494652e-05,
+      "loss": 0.3475,
+      "step": 1785
+    },
+    {
+      "epoch": 1.6682199440820131,
+      "grad_norm": 0.31250485504527636,
+      "learning_rate": 2.466344494304453e-05,
+      "loss": 0.3472,
+      "step": 1790
+    },
+    {
+      "epoch": 1.672879776328052,
+      "grad_norm": 0.28326350593918304,
+      "learning_rate": 2.457714877459441e-05,
+      "loss": 0.3393,
+      "step": 1795
+    },
+    {
+      "epoch": 1.6775396085740915,
+      "grad_norm": 0.34376719631262564,
+      "learning_rate": 2.4490852606144287e-05,
+      "loss": 0.3395,
+      "step": 1800
+    },
+    {
+      "epoch": 1.6821994408201304,
+      "grad_norm": 0.337407219640378,
+      "learning_rate": 2.440455643769417e-05,
+      "loss": 0.3426,
+      "step": 1805
+    },
+    {
+      "epoch": 1.6868592730661698,
+      "grad_norm": 0.284410404461632,
+      "learning_rate": 2.4318260269244047e-05,
+      "loss": 0.3461,
+      "step": 1810
+    },
+    {
+      "epoch": 1.6915191053122087,
+      "grad_norm": 0.2582585348386067,
+      "learning_rate": 2.4231964100793925e-05,
+      "loss": 0.3338,
+      "step": 1815
+    },
+    {
+      "epoch": 1.696178937558248,
+      "grad_norm": 0.2835612157949495,
+      "learning_rate": 2.4145667932343803e-05,
+      "loss": 0.3452,
+      "step": 1820
+    },
+    {
+      "epoch": 1.700838769804287,
+      "grad_norm": 0.26711876772223536,
+      "learning_rate": 2.4059371763893685e-05,
+      "loss": 0.3498,
+      "step": 1825
+    },
+    {
+      "epoch": 1.7054986020503262,
+      "grad_norm": 0.2753356035744131,
+      "learning_rate": 2.3973075595443563e-05,
+      "loss": 0.3548,
+      "step": 1830
+    },
+    {
+      "epoch": 1.7101584342963654,
+      "grad_norm": 0.37119872214335053,
+      "learning_rate": 2.388677942699344e-05,
+      "loss": 0.3498,
+      "step": 1835
+    },
+    {
+      "epoch": 1.7148182665424043,
+      "grad_norm": 0.30893699958152193,
+      "learning_rate": 2.3800483258543323e-05,
+      "loss": 0.346,
+      "step": 1840
+    },
+    {
+      "epoch": 1.7194780987884437,
+      "grad_norm": 0.2709694108627688,
+      "learning_rate": 2.37141870900932e-05,
+      "loss": 0.3505,
+      "step": 1845
+    },
+    {
+      "epoch": 1.7241379310344827,
+      "grad_norm": 0.2723143829630325,
+      "learning_rate": 2.3627890921643083e-05,
+      "loss": 0.3333,
+      "step": 1850
+    },
+    {
+      "epoch": 1.728797763280522,
+      "grad_norm": 0.312759023168839,
+      "learning_rate": 2.3541594753192957e-05,
+      "loss": 0.3432,
+      "step": 1855
+    },
+    {
+      "epoch": 1.733457595526561,
+      "grad_norm": 0.3139145212017304,
+      "learning_rate": 2.345529858474284e-05,
+      "loss": 0.3437,
+      "step": 1860
+    },
+    {
+      "epoch": 1.7381174277726001,
+      "grad_norm": 0.3403229129858024,
+      "learning_rate": 2.3369002416292717e-05,
+      "loss": 0.343,
+      "step": 1865
+    },
+    {
+      "epoch": 1.7427772600186393,
+      "grad_norm": 0.3488122581299756,
+      "learning_rate": 2.32827062478426e-05,
+      "loss": 0.342,
+      "step": 1870
+    },
+    {
+      "epoch": 1.7474370922646785,
+      "grad_norm": 0.276028106726438,
+      "learning_rate": 2.3196410079392474e-05,
+      "loss": 0.3508,
+      "step": 1875
+    },
+    {
+      "epoch": 1.7520969245107176,
+      "grad_norm": 0.34777657040460414,
+      "learning_rate": 2.3110113910942355e-05,
+      "loss": 0.3453,
+      "step": 1880
+    },
+    {
+      "epoch": 1.7567567567567568,
+      "grad_norm": 0.28277033351401737,
+      "learning_rate": 2.3023817742492234e-05,
+      "loss": 0.3506,
+      "step": 1885
+    },
+    {
+      "epoch": 1.761416589002796,
+      "grad_norm": 0.28150023341385944,
+      "learning_rate": 2.2937521574042115e-05,
+      "loss": 0.3481,
+      "step": 1890
+    },
+    {
+      "epoch": 1.766076421248835,
+      "grad_norm": 0.33815617945283766,
+      "learning_rate": 2.285122540559199e-05,
+      "loss": 0.346,
+      "step": 1895
+    },
+    {
+      "epoch": 1.7707362534948743,
+      "grad_norm": 0.5242107536520081,
+      "learning_rate": 2.276492923714187e-05,
+      "loss": 0.3395,
+      "step": 1900
+    },
+    {
+      "epoch": 1.7753960857409132,
+      "grad_norm": 0.27658505149973067,
+      "learning_rate": 2.267863306869175e-05,
+      "loss": 0.3375,
+      "step": 1905
+    },
+    {
+      "epoch": 1.7800559179869526,
+      "grad_norm": 0.26293768037163656,
+      "learning_rate": 2.259233690024163e-05,
+      "loss": 0.3501,
+      "step": 1910
+    },
+    {
+      "epoch": 1.7847157502329916,
+      "grad_norm": 0.2979119949199864,
+      "learning_rate": 2.250604073179151e-05,
+      "loss": 0.3336,
+      "step": 1915
+    },
+    {
+      "epoch": 1.7893755824790307,
+      "grad_norm": 0.32539728714049454,
+      "learning_rate": 2.2419744563341388e-05,
+      "loss": 0.3496,
+      "step": 1920
+    },
+    {
+      "epoch": 1.7940354147250699,
+      "grad_norm": 0.26241094728133707,
+      "learning_rate": 2.233344839489127e-05,
+      "loss": 0.3461,
+      "step": 1925
+    },
+    {
+      "epoch": 1.798695246971109,
+      "grad_norm": 0.29609284549595116,
+      "learning_rate": 2.2247152226441148e-05,
+      "loss": 0.3427,
+      "step": 1930
+    },
+    {
+      "epoch": 1.8033550792171482,
+      "grad_norm": 0.29044024448406247,
+      "learning_rate": 2.2160856057991026e-05,
+      "loss": 0.3477,
+      "step": 1935
+    },
+    {
+      "epoch": 1.8080149114631874,
+      "grad_norm": 0.2692858432041334,
+      "learning_rate": 2.2074559889540904e-05,
+      "loss": 0.3354,
+      "step": 1940
+    },
+    {
+      "epoch": 1.8126747437092265,
+      "grad_norm": 0.27100028935101467,
+      "learning_rate": 2.1988263721090786e-05,
+      "loss": 0.3522,
+      "step": 1945
+    },
+    {
+      "epoch": 1.8173345759552655,
+      "grad_norm": 0.2532552030052246,
+      "learning_rate": 2.1901967552640664e-05,
+      "loss": 0.3354,
+      "step": 1950
+    },
+    {
+      "epoch": 1.8219944082013049,
+      "grad_norm": 0.314442561376392,
+      "learning_rate": 2.1815671384190542e-05,
+      "loss": 0.3435,
+      "step": 1955
+    },
+    {
+      "epoch": 1.8266542404473438,
+      "grad_norm": 0.31415208903879177,
+      "learning_rate": 2.172937521574042e-05,
+      "loss": 0.3472,
+      "step": 1960
+    },
+    {
+      "epoch": 1.8313140726933832,
+      "grad_norm": 0.29355246839128013,
+      "learning_rate": 2.1643079047290302e-05,
+      "loss": 0.3486,
+      "step": 1965
+    },
+    {
+      "epoch": 1.8359739049394221,
+      "grad_norm": 0.28984756640134973,
+      "learning_rate": 2.155678287884018e-05,
+      "loss": 0.3462,
+      "step": 1970
+    },
+    {
+      "epoch": 1.8406337371854613,
+      "grad_norm": 0.2753419738661481,
+      "learning_rate": 2.147048671039006e-05,
+      "loss": 0.3541,
+      "step": 1975
+    },
+    {
+      "epoch": 1.8452935694315005,
+      "grad_norm": 0.2638737876936191,
+      "learning_rate": 2.1384190541939937e-05,
+      "loss": 0.3445,
+      "step": 1980
+    },
+    {
+      "epoch": 1.8499534016775396,
+      "grad_norm": 0.262592387958415,
+      "learning_rate": 2.129789437348982e-05,
+      "loss": 0.3402,
+      "step": 1985
+    },
+    {
+      "epoch": 1.8546132339235788,
+      "grad_norm": 0.26273462103083123,
+      "learning_rate": 2.1211598205039697e-05,
+      "loss": 0.359,
+      "step": 1990
+    },
+    {
+      "epoch": 1.8592730661696177,
+      "grad_norm": 0.3053599234753911,
+      "learning_rate": 2.1125302036589575e-05,
+      "loss": 0.3447,
+      "step": 1995
+    },
+    {
+      "epoch": 1.8639328984156571,
+      "grad_norm": 0.3059994674979238,
+      "learning_rate": 2.1039005868139457e-05,
+      "loss": 0.3435,
+      "step": 2000
+    },
+    {
+      "epoch": 1.868592730661696,
+      "grad_norm": 0.2607025246378089,
+      "learning_rate": 2.0952709699689335e-05,
+      "loss": 0.3373,
+      "step": 2005
+    },
+    {
+      "epoch": 1.8732525629077355,
+      "grad_norm": 0.2861430471124999,
+      "learning_rate": 2.0866413531239216e-05,
+      "loss": 0.3526,
+      "step": 2010
+    },
+    {
+      "epoch": 1.8779123951537744,
+      "grad_norm": 0.2763097244154725,
+      "learning_rate": 2.078011736278909e-05,
+      "loss": 0.3385,
+      "step": 2015
+    },
+    {
+      "epoch": 1.8825722273998136,
+      "grad_norm": 0.2799447185424882,
+      "learning_rate": 2.0693821194338973e-05,
+      "loss": 0.3336,
+      "step": 2020
+    },
+    {
+      "epoch": 1.8872320596458527,
+      "grad_norm": 0.29217744619685787,
+      "learning_rate": 2.060752502588885e-05,
+      "loss": 0.3397,
+      "step": 2025
+    },
+    {
+      "epoch": 1.8918918918918919,
+      "grad_norm": 0.2369462013013538,
+      "learning_rate": 2.0521228857438733e-05,
+      "loss": 0.3465,
+      "step": 2030
+    },
+    {
+      "epoch": 1.896551724137931,
+      "grad_norm": 0.2606923519015433,
+      "learning_rate": 2.0434932688988608e-05,
+      "loss": 0.3398,
+      "step": 2035
+    },
+    {
+      "epoch": 1.9012115563839702,
+      "grad_norm": 0.24433561937288362,
+      "learning_rate": 2.034863652053849e-05,
+      "loss": 0.3426,
+      "step": 2040
+    },
+    {
+      "epoch": 1.9058713886300094,
+      "grad_norm": 0.24027186205793663,
+      "learning_rate": 2.0262340352088367e-05,
+      "loss": 0.3583,
+      "step": 2045
+    },
+    {
+      "epoch": 1.9105312208760483,
+      "grad_norm": 0.2963338411565495,
+      "learning_rate": 2.017604418363825e-05,
+      "loss": 0.3487,
+      "step": 2050
+    },
+    {
+      "epoch": 1.9151910531220877,
+      "grad_norm": 0.30407493721122836,
+      "learning_rate": 2.0089748015188127e-05,
+      "loss": 0.3431,
+      "step": 2055
+    },
+    {
+      "epoch": 1.9198508853681266,
+      "grad_norm": 0.25578502398702385,
+      "learning_rate": 2.0003451846738005e-05,
+      "loss": 0.3509,
+      "step": 2060
+    },
+    {
+      "epoch": 1.924510717614166,
+      "grad_norm": 0.29666053905207496,
+      "learning_rate": 1.9917155678287887e-05,
+      "loss": 0.3402,
+      "step": 2065
+    },
+    {
+      "epoch": 1.929170549860205,
+      "grad_norm": 0.256717876732163,
+      "learning_rate": 1.9830859509837765e-05,
+      "loss": 0.3474,
+      "step": 2070
+    },
+    {
+      "epoch": 1.9338303821062441,
+      "grad_norm": 0.2583678936422422,
+      "learning_rate": 1.9744563341387643e-05,
+      "loss": 0.3491,
+      "step": 2075
+    },
+    {
+      "epoch": 1.9384902143522833,
+      "grad_norm": 0.2950508942608553,
+      "learning_rate": 1.9658267172937522e-05,
+      "loss": 0.3441,
+      "step": 2080
+    },
+    {
+      "epoch": 1.9431500465983225,
+      "grad_norm": 0.36955576638078624,
+      "learning_rate": 1.9571971004487403e-05,
+      "loss": 0.3438,
+      "step": 2085
+    },
+    {
+      "epoch": 1.9478098788443616,
+      "grad_norm": 0.3035648067236326,
+      "learning_rate": 1.948567483603728e-05,
+      "loss": 0.3498,
+      "step": 2090
+    },
+    {
+      "epoch": 1.9524697110904008,
+      "grad_norm": 0.3134746870815898,
+      "learning_rate": 1.939937866758716e-05,
+      "loss": 0.3321,
+      "step": 2095
+    },
+    {
+      "epoch": 1.95712954333644,
+      "grad_norm": 0.33205293171672917,
+      "learning_rate": 1.9313082499137038e-05,
+      "loss": 0.3493,
+      "step": 2100
+    },
+    {
+      "epoch": 1.961789375582479,
+      "grad_norm": 0.3166250804073325,
+      "learning_rate": 1.922678633068692e-05,
+      "loss": 0.3439,
+      "step": 2105
+    },
+    {
+      "epoch": 1.9664492078285183,
+      "grad_norm": 0.2735011532790032,
+      "learning_rate": 1.9140490162236798e-05,
+      "loss": 0.3493,
+      "step": 2110
+    },
+    {
+      "epoch": 1.9711090400745572,
+      "grad_norm": 0.27483640823140315,
+      "learning_rate": 1.9054193993786676e-05,
+      "loss": 0.3565,
+      "step": 2115
+    },
+    {
+      "epoch": 1.9757688723205966,
+      "grad_norm": 0.2528402059369525,
+      "learning_rate": 1.8967897825336554e-05,
+      "loss": 0.3444,
+      "step": 2120
+    },
+    {
+      "epoch": 1.9804287045666356,
+      "grad_norm": 0.24447651503450293,
+      "learning_rate": 1.8881601656886436e-05,
+      "loss": 0.3418,
+      "step": 2125
+    },
+    {
+      "epoch": 1.9850885368126747,
+      "grad_norm": 0.2661148510991396,
+      "learning_rate": 1.8795305488436314e-05,
+      "loss": 0.3458,
+      "step": 2130
+    },
+    {
+      "epoch": 1.9897483690587139,
+      "grad_norm": 0.3215277813688595,
+      "learning_rate": 1.8709009319986192e-05,
+      "loss": 0.3342,
+      "step": 2135
+    },
+    {
+      "epoch": 1.994408201304753,
+      "grad_norm": 0.28210057507380515,
+      "learning_rate": 1.8622713151536074e-05,
+      "loss": 0.3411,
+      "step": 2140
+    },
+    {
+      "epoch": 1.9990680335507922,
+      "grad_norm": 0.24421351565362545,
+      "learning_rate": 1.8536416983085952e-05,
+      "loss": 0.3459,
+      "step": 2145
+    },
+    {
+      "epoch": 2.003727865796831,
+      "grad_norm": 0.27803726882878427,
+      "learning_rate": 1.8450120814635834e-05,
+      "loss": 0.284,
+      "step": 2150
+    },
+    {
+      "epoch": 2.0083876980428705,
+      "grad_norm": 0.289132512553544,
+      "learning_rate": 1.836382464618571e-05,
+      "loss": 0.2864,
+      "step": 2155
+    },
+    {
+      "epoch": 2.0130475302889095,
+      "grad_norm": 0.2620365826195705,
+      "learning_rate": 1.827752847773559e-05,
+      "loss": 0.2802,
+      "step": 2160
+    },
+    {
+      "epoch": 2.017707362534949,
+      "grad_norm": 0.267513690683278,
+      "learning_rate": 1.819123230928547e-05,
+      "loss": 0.2827,
+      "step": 2165
+    },
+    {
+      "epoch": 2.022367194780988,
+      "grad_norm": 0.24650696041571432,
+      "learning_rate": 1.810493614083535e-05,
+      "loss": 0.282,
+      "step": 2170
+    },
+    {
+      "epoch": 2.027027027027027,
+      "grad_norm": 0.23594499949396758,
+      "learning_rate": 1.8018639972385225e-05,
+      "loss": 0.2769,
+      "step": 2175
+    },
+    {
+      "epoch": 2.031686859273066,
+      "grad_norm": 0.25002507657849443,
+      "learning_rate": 1.7932343803935107e-05,
+      "loss": 0.2817,
+      "step": 2180
+    },
+    {
+      "epoch": 2.0363466915191055,
+      "grad_norm": 0.2421805814272178,
+      "learning_rate": 1.7846047635484985e-05,
+      "loss": 0.2781,
+      "step": 2185
+    },
+    {
+      "epoch": 2.0410065237651445,
+      "grad_norm": 0.2446711558000852,
+      "learning_rate": 1.7759751467034866e-05,
+      "loss": 0.2782,
+      "step": 2190
+    },
+    {
+      "epoch": 2.0456663560111834,
+      "grad_norm": 0.2893143893353549,
+      "learning_rate": 1.767345529858474e-05,
+      "loss": 0.2881,
+      "step": 2195
+    },
+    {
+      "epoch": 2.050326188257223,
+      "grad_norm": 0.2361572137636827,
+      "learning_rate": 1.7587159130134623e-05,
+      "loss": 0.2879,
+      "step": 2200
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 3219,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 550,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.8830647294511022e+18,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-2200/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:813a77be1c465ef5e10e4edc1f2885fd2a8dd215211b3e931e4d041d82d8bcec
+size 7480

checkpoint-2200/zero_to_fp32.py

@@ -0,0 +1,674 @@
+#!/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 json
+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)
+
+        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 files:
+        state_dict = torch.load(f, map_location=device)
+        # 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}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) 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")
+
+
+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 = fp32_flat_groups[0].numel() * 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
+    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}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        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 get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=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
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    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.
+
+    """
+    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")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+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)
+
+    # 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")
+        state_dict_split = split_torch_state_dict_into_shards(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
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # 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)