Training in progress, epoch 2, checkpoint

checkpoint-2148/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-2148/config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "Qwen7B-HP-AMP",
+  "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": 1000000.0,
+  "sliding_window": 131072,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

checkpoint-2148/generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.49.0"
+}

checkpoint-2148/global_step2147/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:a14daf30692502ce252c3f6c4e6e9c96e13063c40997518152cf68670a06154b
+size 12117257612

checkpoint-2148/global_step2147/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:d6cf7676fd4a0a1c0c39ce8885b957d1ade8d6f3d30da90a92f18b759c7ec459
+size 12117257612

checkpoint-2148/global_step2147/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:d83b166b6ca9a9f85c0fc497d4f1d6b8071a315f9a930eb723a94ea9de937605
+size 12117257612

checkpoint-2148/global_step2147/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:e4ab55f44d5b66bc9e37586e4bd27b237fd231f188e81bcf8dad197429625700
+size 12117257612

checkpoint-2148/global_step2147/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84e4d52bbdeee3c4f0be515e84a02773f186b5118616912b96df9a9df9587d64
+size 12117257612

checkpoint-2148/global_step2147/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:488778e4b68f24b47595f60dedb597e82c822f1c9c29a3061869bf4536aed1c8
+size 12117257612

checkpoint-2148/global_step2147/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:765f05b468e6a3e8bae4fe3356cf92d556ba38fb1ed18cae6fc9d5f0c51a4b75
+size 12117257612

checkpoint-2148/global_step2147/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cb186d2204b2751307c769bd799a9d21c69497050a49f931bd093142f539080
+size 12117257612

checkpoint-2148/global_step2147/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac16b658db01c302634ccaf139eb49471821ce0eb58eda48f92dfc0bf6d6a283
+size 256905

checkpoint-2148/global_step2147/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b679a7a8adb74687d574da41a488c5180c991df681e6f5aeff97fba683ccec8
+size 256905

checkpoint-2148/global_step2147/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbfb8c014e9620d18aee8d7add1593439632ec8e9ea9866c4cde80096f46bed9
+size 256905

checkpoint-2148/global_step2147/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc41e6276ec67247b79af681021e39e7d368cdef5df0c097007c3a9b82295b61
+size 256905

checkpoint-2148/global_step2147/zero_pp_rank_4_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9296d159e43f551899e0010470cac82d9f99058ab6ac4134d583a88b4475b7df
+size 256905

checkpoint-2148/global_step2147/zero_pp_rank_5_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16885cd3cd98ae22ff3a19415d5211753288d4f662c973be46cc3c7a7e17d029
+size 256905

checkpoint-2148/global_step2147/zero_pp_rank_6_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a3ca1b330100f1cfb8923af5c53dd6402373f2a68a2e36cc4d0c13f2634f012
+size 256905

checkpoint-2148/global_step2147/zero_pp_rank_7_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c401b6b35dfcc48e28491d7818dcd2cb1c266f0994727e8153909f1d135c90eb
+size 256905

checkpoint-2148/latest

@@ -0,0 +1 @@
+global_step2147

checkpoint-2148/model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f42241f6128380cb284363f9040077a885bcd14611e318b4b2d0d05175f65a9
+size 4947447072

checkpoint-2148/model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19d963e5fc4f412d1a266ac8097800ae2ff396a16233fdf20617bc2402601ae9
+size 4991571912

checkpoint-2148/model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f88ea8f8c052cf813720fe7e1f763a4fd33893bd10bca7cb5c88be90e9f4d569
+size 4991571984

checkpoint-2148/model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0fbb2ab89a491cd0720ab931baa1c8e23d5fe4f8299d6f4b3b5802c6fc7e0e0
+size 1225807416

checkpoint-2148/model.safetensors.index.json

@@ -0,0 +1,542 @@
+{
+  "metadata": {
+    "total_size": 16156336672
+  },
+  "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.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.amp_v_proj.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-00003-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00003-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-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.amp_v_proj.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-00003-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.amp_v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00003-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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.amp_v_proj.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-00004-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00004-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-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.amp_k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.amp_k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.amp_q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.amp_q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.amp_scaler": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.amp_v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.amp_v_proj.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-00002-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00002-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-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.amp_k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.amp_k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.amp_q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.amp_q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.amp_scaler": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.amp_v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.amp_v_proj.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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.amp_v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00002-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.amp_k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.amp_k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.amp_q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.amp_q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.amp_scaler": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.amp_v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.amp_v_proj.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-00004-of-00004.safetensors"
+  }
+}

checkpoint-2148/rng_state_0.pth

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

checkpoint-2148/rng_state_1.pth

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

checkpoint-2148/rng_state_2.pth

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

checkpoint-2148/rng_state_3.pth

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

checkpoint-2148/rng_state_4.pth

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

checkpoint-2148/rng_state_5.pth

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

checkpoint-2148/rng_state_6.pth

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

checkpoint-2148/rng_state_7.pth

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

checkpoint-2148/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56570f0b9c4a46ef6adafa222772f59b17d6eaad0a5bc9228f66bdf0ffeeffc0
+size 1064

checkpoint-2148/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-2148/tokenizer.json

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

checkpoint-2148/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        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\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\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|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-2148/trainer_state.json

@@ -0,0 +1,3036 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 2148,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.004657661853749418,
+      "grad_norm": 27.66617259818207,
+      "learning_rate": 5.813953488372093e-07,
+      "loss": 1.2864,
+      "step": 5
+    },
+    {
+      "epoch": 0.009315323707498836,
+      "grad_norm": 19.626216359351606,
+      "learning_rate": 1.1627906976744186e-06,
+      "loss": 1.2277,
+      "step": 10
+    },
+    {
+      "epoch": 0.013972985561248253,
+      "grad_norm": 7.689899823367007,
+      "learning_rate": 1.744186046511628e-06,
+      "loss": 1.0529,
+      "step": 15
+    },
+    {
+      "epoch": 0.018630647414997672,
+      "grad_norm": 4.9632975953671465,
+      "learning_rate": 2.325581395348837e-06,
+      "loss": 0.907,
+      "step": 20
+    },
+    {
+      "epoch": 0.02328830926874709,
+      "grad_norm": 2.000237294817273,
+      "learning_rate": 2.9069767441860468e-06,
+      "loss": 0.7884,
+      "step": 25
+    },
+    {
+      "epoch": 0.027945971122496506,
+      "grad_norm": 1.3650044798865562,
+      "learning_rate": 3.488372093023256e-06,
+      "loss": 0.7091,
+      "step": 30
+    },
+    {
+      "epoch": 0.032603632976245925,
+      "grad_norm": 1.0145438785392298,
+      "learning_rate": 4.0697674418604655e-06,
+      "loss": 0.6666,
+      "step": 35
+    },
+    {
+      "epoch": 0.037261294829995344,
+      "grad_norm": 0.7666883370082983,
+      "learning_rate": 4.651162790697674e-06,
+      "loss": 0.6289,
+      "step": 40
+    },
+    {
+      "epoch": 0.04191895668374476,
+      "grad_norm": 0.6977041513064055,
+      "learning_rate": 5.232558139534884e-06,
+      "loss": 0.5772,
+      "step": 45
+    },
+    {
+      "epoch": 0.04657661853749418,
+      "grad_norm": 0.6177135531852819,
+      "learning_rate": 5.8139534883720935e-06,
+      "loss": 0.5729,
+      "step": 50
+    },
+    {
+      "epoch": 0.05123428039124359,
+      "grad_norm": 0.5909988418524453,
+      "learning_rate": 6.395348837209303e-06,
+      "loss": 0.5555,
+      "step": 55
+    },
+    {
+      "epoch": 0.05589194224499301,
+      "grad_norm": 0.6995185567241584,
+      "learning_rate": 6.976744186046512e-06,
+      "loss": 0.5472,
+      "step": 60
+    },
+    {
+      "epoch": 0.06054960409874243,
+      "grad_norm": 0.7744854841742352,
+      "learning_rate": 7.558139534883721e-06,
+      "loss": 0.5379,
+      "step": 65
+    },
+    {
+      "epoch": 0.06520726595249185,
+      "grad_norm": 0.5473128068676195,
+      "learning_rate": 8.139534883720931e-06,
+      "loss": 0.5149,
+      "step": 70
+    },
+    {
+      "epoch": 0.06986492780624126,
+      "grad_norm": 0.711636506166052,
+      "learning_rate": 8.72093023255814e-06,
+      "loss": 0.5151,
+      "step": 75
+    },
+    {
+      "epoch": 0.07452258965999069,
+      "grad_norm": 0.6813906468319959,
+      "learning_rate": 9.302325581395349e-06,
+      "loss": 0.5252,
+      "step": 80
+    },
+    {
+      "epoch": 0.0791802515137401,
+      "grad_norm": 0.7134287439415516,
+      "learning_rate": 9.883720930232558e-06,
+      "loss": 0.5024,
+      "step": 85
+    },
+    {
+      "epoch": 0.08383791336748952,
+      "grad_norm": 0.6992933648203333,
+      "learning_rate": 1.0465116279069768e-05,
+      "loss": 0.5085,
+      "step": 90
+    },
+    {
+      "epoch": 0.08849557522123894,
+      "grad_norm": 0.6462820616067007,
+      "learning_rate": 1.1046511627906977e-05,
+      "loss": 0.5129,
+      "step": 95
+    },
+    {
+      "epoch": 0.09315323707498836,
+      "grad_norm": 0.67397577790992,
+      "learning_rate": 1.1627906976744187e-05,
+      "loss": 0.5034,
+      "step": 100
+    },
+    {
+      "epoch": 0.09781089892873777,
+      "grad_norm": 0.7146380645525483,
+      "learning_rate": 1.2209302325581395e-05,
+      "loss": 0.506,
+      "step": 105
+    },
+    {
+      "epoch": 0.10246856078248719,
+      "grad_norm": 0.7569877075727182,
+      "learning_rate": 1.2790697674418606e-05,
+      "loss": 0.4939,
+      "step": 110
+    },
+    {
+      "epoch": 0.10712622263623661,
+      "grad_norm": 0.5765767741771239,
+      "learning_rate": 1.3372093023255814e-05,
+      "loss": 0.4932,
+      "step": 115
+    },
+    {
+      "epoch": 0.11178388448998602,
+      "grad_norm": 0.7689970599513694,
+      "learning_rate": 1.3953488372093024e-05,
+      "loss": 0.5041,
+      "step": 120
+    },
+    {
+      "epoch": 0.11644154634373545,
+      "grad_norm": 0.7710814056961762,
+      "learning_rate": 1.4534883720930233e-05,
+      "loss": 0.4846,
+      "step": 125
+    },
+    {
+      "epoch": 0.12109920819748486,
+      "grad_norm": 0.8067799411588391,
+      "learning_rate": 1.5116279069767441e-05,
+      "loss": 0.4868,
+      "step": 130
+    },
+    {
+      "epoch": 0.1257568700512343,
+      "grad_norm": 0.7337587312339595,
+      "learning_rate": 1.569767441860465e-05,
+      "loss": 0.4937,
+      "step": 135
+    },
+    {
+      "epoch": 0.1304145319049837,
+      "grad_norm": 0.8156553975576142,
+      "learning_rate": 1.6279069767441862e-05,
+      "loss": 0.4871,
+      "step": 140
+    },
+    {
+      "epoch": 0.1350721937587331,
+      "grad_norm": 0.9350450002033659,
+      "learning_rate": 1.686046511627907e-05,
+      "loss": 0.4972,
+      "step": 145
+    },
+    {
+      "epoch": 0.13972985561248252,
+      "grad_norm": 0.7681334266478443,
+      "learning_rate": 1.744186046511628e-05,
+      "loss": 0.4907,
+      "step": 150
+    },
+    {
+      "epoch": 0.14438751746623196,
+      "grad_norm": 0.84555038297673,
+      "learning_rate": 1.802325581395349e-05,
+      "loss": 0.4723,
+      "step": 155
+    },
+    {
+      "epoch": 0.14904517931998137,
+      "grad_norm": 0.8372039362868026,
+      "learning_rate": 1.8604651162790697e-05,
+      "loss": 0.4793,
+      "step": 160
+    },
+    {
+      "epoch": 0.1537028411737308,
+      "grad_norm": 0.7817304569895932,
+      "learning_rate": 1.918604651162791e-05,
+      "loss": 0.4798,
+      "step": 165
+    },
+    {
+      "epoch": 0.1583605030274802,
+      "grad_norm": 0.9789922595882735,
+      "learning_rate": 1.9767441860465116e-05,
+      "loss": 0.4783,
+      "step": 170
+    },
+    {
+      "epoch": 0.1630181648812296,
+      "grad_norm": 0.7335091977994727,
+      "learning_rate": 2.0348837209302328e-05,
+      "loss": 0.4893,
+      "step": 175
+    },
+    {
+      "epoch": 0.16767582673497905,
+      "grad_norm": 0.8155578488051264,
+      "learning_rate": 2.0930232558139536e-05,
+      "loss": 0.484,
+      "step": 180
+    },
+    {
+      "epoch": 0.17233348858872846,
+      "grad_norm": 0.8144677339935327,
+      "learning_rate": 2.1511627906976744e-05,
+      "loss": 0.482,
+      "step": 185
+    },
+    {
+      "epoch": 0.17699115044247787,
+      "grad_norm": 0.8341691871989111,
+      "learning_rate": 2.2093023255813955e-05,
+      "loss": 0.4753,
+      "step": 190
+    },
+    {
+      "epoch": 0.18164881229622729,
+      "grad_norm": 0.7550827853742104,
+      "learning_rate": 2.2674418604651163e-05,
+      "loss": 0.4811,
+      "step": 195
+    },
+    {
+      "epoch": 0.18630647414997673,
+      "grad_norm": 0.726150592947813,
+      "learning_rate": 2.3255813953488374e-05,
+      "loss": 0.4548,
+      "step": 200
+    },
+    {
+      "epoch": 0.19096413600372614,
+      "grad_norm": 0.9618454710454514,
+      "learning_rate": 2.3837209302325582e-05,
+      "loss": 0.4765,
+      "step": 205
+    },
+    {
+      "epoch": 0.19562179785747555,
+      "grad_norm": 0.8357087005031101,
+      "learning_rate": 2.441860465116279e-05,
+      "loss": 0.4664,
+      "step": 210
+    },
+    {
+      "epoch": 0.20027945971122496,
+      "grad_norm": 0.9769412255461828,
+      "learning_rate": 2.5e-05,
+      "loss": 0.4772,
+      "step": 215
+    },
+    {
+      "epoch": 0.20493712156497437,
+      "grad_norm": 0.819002974551843,
+      "learning_rate": 2.5581395348837212e-05,
+      "loss": 0.4731,
+      "step": 220
+    },
+    {
+      "epoch": 0.2095947834187238,
+      "grad_norm": 1.0374332843969527,
+      "learning_rate": 2.616279069767442e-05,
+      "loss": 0.4763,
+      "step": 225
+    },
+    {
+      "epoch": 0.21425244527247322,
+      "grad_norm": 0.8985668587375348,
+      "learning_rate": 2.674418604651163e-05,
+      "loss": 0.4745,
+      "step": 230
+    },
+    {
+      "epoch": 0.21891010712622264,
+      "grad_norm": 1.0814445549381904,
+      "learning_rate": 2.7325581395348836e-05,
+      "loss": 0.4695,
+      "step": 235
+    },
+    {
+      "epoch": 0.22356776897997205,
+      "grad_norm": 1.0714624697860875,
+      "learning_rate": 2.7906976744186048e-05,
+      "loss": 0.4687,
+      "step": 240
+    },
+    {
+      "epoch": 0.22822543083372146,
+      "grad_norm": 0.9839374258594388,
+      "learning_rate": 2.848837209302326e-05,
+      "loss": 0.4598,
+      "step": 245
+    },
+    {
+      "epoch": 0.2328830926874709,
+      "grad_norm": 0.9846036088169035,
+      "learning_rate": 2.9069767441860467e-05,
+      "loss": 0.4569,
+      "step": 250
+    },
+    {
+      "epoch": 0.2375407545412203,
+      "grad_norm": 0.7487150924303477,
+      "learning_rate": 2.9651162790697678e-05,
+      "loss": 0.4676,
+      "step": 255
+    },
+    {
+      "epoch": 0.24219841639496972,
+      "grad_norm": 0.8226804932307876,
+      "learning_rate": 3.0232558139534883e-05,
+      "loss": 0.4722,
+      "step": 260
+    },
+    {
+      "epoch": 0.24685607824871914,
+      "grad_norm": 0.7711022626726491,
+      "learning_rate": 3.081395348837209e-05,
+      "loss": 0.4658,
+      "step": 265
+    },
+    {
+      "epoch": 0.2515137401024686,
+      "grad_norm": 0.8932122698414526,
+      "learning_rate": 3.13953488372093e-05,
+      "loss": 0.4557,
+      "step": 270
+    },
+    {
+      "epoch": 0.25617140195621796,
+      "grad_norm": 0.7549553477075355,
+      "learning_rate": 3.197674418604651e-05,
+      "loss": 0.4716,
+      "step": 275
+    },
+    {
+      "epoch": 0.2608290638099674,
+      "grad_norm": 1.203597171762952,
+      "learning_rate": 3.2558139534883724e-05,
+      "loss": 0.4602,
+      "step": 280
+    },
+    {
+      "epoch": 0.26548672566371684,
+      "grad_norm": 0.6877770638858097,
+      "learning_rate": 3.313953488372093e-05,
+      "loss": 0.4589,
+      "step": 285
+    },
+    {
+      "epoch": 0.2701443875174662,
+      "grad_norm": 0.9865065100124113,
+      "learning_rate": 3.372093023255814e-05,
+      "loss": 0.4577,
+      "step": 290
+    },
+    {
+      "epoch": 0.27480204937121566,
+      "grad_norm": 0.9375168702189124,
+      "learning_rate": 3.430232558139535e-05,
+      "loss": 0.4516,
+      "step": 295
+    },
+    {
+      "epoch": 0.27945971122496505,
+      "grad_norm": 0.7788845536490606,
+      "learning_rate": 3.488372093023256e-05,
+      "loss": 0.4633,
+      "step": 300
+    },
+    {
+      "epoch": 0.2841173730787145,
+      "grad_norm": 0.8098072656748624,
+      "learning_rate": 3.5465116279069774e-05,
+      "loss": 0.4691,
+      "step": 305
+    },
+    {
+      "epoch": 0.2887750349324639,
+      "grad_norm": 1.3633357913402282,
+      "learning_rate": 3.604651162790698e-05,
+      "loss": 0.4662,
+      "step": 310
+    },
+    {
+      "epoch": 0.2934326967862133,
+      "grad_norm": 0.8073757027557402,
+      "learning_rate": 3.662790697674418e-05,
+      "loss": 0.4551,
+      "step": 315
+    },
+    {
+      "epoch": 0.29809035863996275,
+      "grad_norm": 1.0699898191361614,
+      "learning_rate": 3.7209302325581394e-05,
+      "loss": 0.4614,
+      "step": 320
+    },
+    {
+      "epoch": 0.30274802049371213,
+      "grad_norm": 0.7424537685527064,
+      "learning_rate": 3.7790697674418606e-05,
+      "loss": 0.4637,
+      "step": 325
+    },
+    {
+      "epoch": 0.3074056823474616,
+      "grad_norm": 1.0663211887340074,
+      "learning_rate": 3.837209302325582e-05,
+      "loss": 0.463,
+      "step": 330
+    },
+    {
+      "epoch": 0.312063344201211,
+      "grad_norm": 0.9616713033039357,
+      "learning_rate": 3.895348837209303e-05,
+      "loss": 0.4483,
+      "step": 335
+    },
+    {
+      "epoch": 0.3167210060549604,
+      "grad_norm": 0.9289014338139833,
+      "learning_rate": 3.953488372093023e-05,
+      "loss": 0.4618,
+      "step": 340
+    },
+    {
+      "epoch": 0.32137866790870984,
+      "grad_norm": 0.9205940376448726,
+      "learning_rate": 4.0116279069767444e-05,
+      "loss": 0.4516,
+      "step": 345
+    },
+    {
+      "epoch": 0.3260363297624592,
+      "grad_norm": 0.9475824983358404,
+      "learning_rate": 4.0697674418604655e-05,
+      "loss": 0.4545,
+      "step": 350
+    },
+    {
+      "epoch": 0.33069399161620866,
+      "grad_norm": 1.315055782839884,
+      "learning_rate": 4.127906976744187e-05,
+      "loss": 0.452,
+      "step": 355
+    },
+    {
+      "epoch": 0.3353516534699581,
+      "grad_norm": 1.068696802277167,
+      "learning_rate": 4.186046511627907e-05,
+      "loss": 0.4625,
+      "step": 360
+    },
+    {
+      "epoch": 0.3400093153237075,
+      "grad_norm": 1.036508382207286,
+      "learning_rate": 4.2441860465116276e-05,
+      "loss": 0.4615,
+      "step": 365
+    },
+    {
+      "epoch": 0.3446669771774569,
+      "grad_norm": 0.9771206427059291,
+      "learning_rate": 4.302325581395349e-05,
+      "loss": 0.472,
+      "step": 370
+    },
+    {
+      "epoch": 0.3493246390312063,
+      "grad_norm": 0.9364995309041826,
+      "learning_rate": 4.36046511627907e-05,
+      "loss": 0.4662,
+      "step": 375
+    },
+    {
+      "epoch": 0.35398230088495575,
+      "grad_norm": 0.8666798423814609,
+      "learning_rate": 4.418604651162791e-05,
+      "loss": 0.4513,
+      "step": 380
+    },
+    {
+      "epoch": 0.3586399627387052,
+      "grad_norm": 0.6625516854940565,
+      "learning_rate": 4.476744186046512e-05,
+      "loss": 0.4458,
+      "step": 385
+    },
+    {
+      "epoch": 0.36329762459245457,
+      "grad_norm": 0.5910994980517921,
+      "learning_rate": 4.5348837209302326e-05,
+      "loss": 0.4518,
+      "step": 390
+    },
+    {
+      "epoch": 0.367955286446204,
+      "grad_norm": 0.8576114090853963,
+      "learning_rate": 4.593023255813954e-05,
+      "loss": 0.4472,
+      "step": 395
+    },
+    {
+      "epoch": 0.37261294829995345,
+      "grad_norm": 1.059751439208873,
+      "learning_rate": 4.651162790697675e-05,
+      "loss": 0.4453,
+      "step": 400
+    },
+    {
+      "epoch": 0.37727061015370283,
+      "grad_norm": 0.7913724415049757,
+      "learning_rate": 4.709302325581396e-05,
+      "loss": 0.4492,
+      "step": 405
+    },
+    {
+      "epoch": 0.3819282720074523,
+      "grad_norm": 0.7471961729660666,
+      "learning_rate": 4.7674418604651164e-05,
+      "loss": 0.4559,
+      "step": 410
+    },
+    {
+      "epoch": 0.38658593386120166,
+      "grad_norm": 0.7938017748454547,
+      "learning_rate": 4.8255813953488375e-05,
+      "loss": 0.4556,
+      "step": 415
+    },
+    {
+      "epoch": 0.3912435957149511,
+      "grad_norm": 0.861398782650763,
+      "learning_rate": 4.883720930232558e-05,
+      "loss": 0.4457,
+      "step": 420
+    },
+    {
+      "epoch": 0.39590125756870054,
+      "grad_norm": 1.164899725237036,
+      "learning_rate": 4.941860465116279e-05,
+      "loss": 0.4619,
+      "step": 425
+    },
+    {
+      "epoch": 0.4005589194224499,
+      "grad_norm": 0.9327630997758065,
+      "learning_rate": 5e-05,
+      "loss": 0.4553,
+      "step": 430
+    },
+    {
+      "epoch": 0.40521658127619936,
+      "grad_norm": 0.7755438754585425,
+      "learning_rate": 4.9935266701191095e-05,
+      "loss": 0.4589,
+      "step": 435
+    },
+    {
+      "epoch": 0.40987424312994875,
+      "grad_norm": 0.6084430372051816,
+      "learning_rate": 4.987053340238219e-05,
+      "loss": 0.4659,
+      "step": 440
+    },
+    {
+      "epoch": 0.4145319049836982,
+      "grad_norm": 0.706361406749313,
+      "learning_rate": 4.980580010357328e-05,
+      "loss": 0.4454,
+      "step": 445
+    },
+    {
+      "epoch": 0.4191895668374476,
+      "grad_norm": 0.8310997319677449,
+      "learning_rate": 4.9741066804764374e-05,
+      "loss": 0.4492,
+      "step": 450
+    },
+    {
+      "epoch": 0.423847228691197,
+      "grad_norm": 0.7992980724203008,
+      "learning_rate": 4.967633350595546e-05,
+      "loss": 0.4677,
+      "step": 455
+    },
+    {
+      "epoch": 0.42850489054494645,
+      "grad_norm": 0.7848440239850348,
+      "learning_rate": 4.961160020714656e-05,
+      "loss": 0.4484,
+      "step": 460
+    },
+    {
+      "epoch": 0.43316255239869583,
+      "grad_norm": 0.6875355874505797,
+      "learning_rate": 4.954686690833765e-05,
+      "loss": 0.4592,
+      "step": 465
+    },
+    {
+      "epoch": 0.43782021425244527,
+      "grad_norm": 0.9019111936374453,
+      "learning_rate": 4.948213360952874e-05,
+      "loss": 0.4507,
+      "step": 470
+    },
+    {
+      "epoch": 0.4424778761061947,
+      "grad_norm": 0.9237443867694993,
+      "learning_rate": 4.941740031071983e-05,
+      "loss": 0.46,
+      "step": 475
+    },
+    {
+      "epoch": 0.4471355379599441,
+      "grad_norm": 0.7227765777311265,
+      "learning_rate": 4.935266701191093e-05,
+      "loss": 0.4448,
+      "step": 480
+    },
+    {
+      "epoch": 0.45179319981369354,
+      "grad_norm": 0.7627976373615327,
+      "learning_rate": 4.9287933713102025e-05,
+      "loss": 0.4467,
+      "step": 485
+    },
+    {
+      "epoch": 0.4564508616674429,
+      "grad_norm": 0.8963050574087497,
+      "learning_rate": 4.922320041429311e-05,
+      "loss": 0.45,
+      "step": 490
+    },
+    {
+      "epoch": 0.46110852352119236,
+      "grad_norm": 0.629147905901097,
+      "learning_rate": 4.915846711548421e-05,
+      "loss": 0.4427,
+      "step": 495
+    },
+    {
+      "epoch": 0.4657661853749418,
+      "grad_norm": 0.5883243359451029,
+      "learning_rate": 4.9093733816675304e-05,
+      "loss": 0.4491,
+      "step": 500
+    },
+    {
+      "epoch": 0.4704238472286912,
+      "grad_norm": 0.5369867711481909,
+      "learning_rate": 4.902900051786639e-05,
+      "loss": 0.4435,
+      "step": 505
+    },
+    {
+      "epoch": 0.4750815090824406,
+      "grad_norm": 0.7877373044239472,
+      "learning_rate": 4.8964267219057483e-05,
+      "loss": 0.4392,
+      "step": 510
+    },
+    {
+      "epoch": 0.47973917093619,
+      "grad_norm": 0.5720363420014942,
+      "learning_rate": 4.889953392024858e-05,
+      "loss": 0.4502,
+      "step": 515
+    },
+    {
+      "epoch": 0.48439683278993945,
+      "grad_norm": 0.802768212629227,
+      "learning_rate": 4.883480062143967e-05,
+      "loss": 0.4479,
+      "step": 520
+    },
+    {
+      "epoch": 0.4890544946436889,
+      "grad_norm": 0.9089354308220536,
+      "learning_rate": 4.877006732263076e-05,
+      "loss": 0.4562,
+      "step": 525
+    },
+    {
+      "epoch": 0.49371215649743827,
+      "grad_norm": 0.9062598882129868,
+      "learning_rate": 4.8705334023821855e-05,
+      "loss": 0.4433,
+      "step": 530
+    },
+    {
+      "epoch": 0.4983698183511877,
+      "grad_norm": 1.0614931500809168,
+      "learning_rate": 4.864060072501295e-05,
+      "loss": 0.4509,
+      "step": 535
+    },
+    {
+      "epoch": 0.5030274802049371,
+      "grad_norm": 0.6656921373680668,
+      "learning_rate": 4.857586742620404e-05,
+      "loss": 0.4526,
+      "step": 540
+    },
+    {
+      "epoch": 0.5076851420586865,
+      "grad_norm": 0.761135804704884,
+      "learning_rate": 4.8511134127395134e-05,
+      "loss": 0.4428,
+      "step": 545
+    },
+    {
+      "epoch": 0.5123428039124359,
+      "grad_norm": 0.7524291911003331,
+      "learning_rate": 4.844640082858623e-05,
+      "loss": 0.4559,
+      "step": 550
+    },
+    {
+      "epoch": 0.5170004657661854,
+      "grad_norm": 0.7643694469846594,
+      "learning_rate": 4.838166752977732e-05,
+      "loss": 0.4441,
+      "step": 555
+    },
+    {
+      "epoch": 0.5216581276199348,
+      "grad_norm": 0.6727074125902812,
+      "learning_rate": 4.831693423096841e-05,
+      "loss": 0.4482,
+      "step": 560
+    },
+    {
+      "epoch": 0.5263157894736842,
+      "grad_norm": 0.6080795466448415,
+      "learning_rate": 4.82522009321595e-05,
+      "loss": 0.4464,
+      "step": 565
+    },
+    {
+      "epoch": 0.5309734513274337,
+      "grad_norm": 0.7154575256811082,
+      "learning_rate": 4.81874676333506e-05,
+      "loss": 0.4456,
+      "step": 570
+    },
+    {
+      "epoch": 0.5356311131811831,
+      "grad_norm": 0.5881145419877987,
+      "learning_rate": 4.812273433454169e-05,
+      "loss": 0.4429,
+      "step": 575
+    },
+    {
+      "epoch": 0.5402887750349324,
+      "grad_norm": 0.5629504481988451,
+      "learning_rate": 4.8058001035732785e-05,
+      "loss": 0.4428,
+      "step": 580
+    },
+    {
+      "epoch": 0.5449464368886818,
+      "grad_norm": 0.6740507784348432,
+      "learning_rate": 4.799326773692387e-05,
+      "loss": 0.4389,
+      "step": 585
+    },
+    {
+      "epoch": 0.5496040987424313,
+      "grad_norm": 0.5590413217515213,
+      "learning_rate": 4.792853443811497e-05,
+      "loss": 0.4543,
+      "step": 590
+    },
+    {
+      "epoch": 0.5542617605961807,
+      "grad_norm": 0.648175325079465,
+      "learning_rate": 4.7863801139306064e-05,
+      "loss": 0.4446,
+      "step": 595
+    },
+    {
+      "epoch": 0.5589194224499301,
+      "grad_norm": 0.6553497135224098,
+      "learning_rate": 4.779906784049715e-05,
+      "loss": 0.4501,
+      "step": 600
+    },
+    {
+      "epoch": 0.5635770843036796,
+      "grad_norm": 0.6606737255970081,
+      "learning_rate": 4.773433454168825e-05,
+      "loss": 0.4256,
+      "step": 605
+    },
+    {
+      "epoch": 0.568234746157429,
+      "grad_norm": 0.7570935534531892,
+      "learning_rate": 4.766960124287934e-05,
+      "loss": 0.4434,
+      "step": 610
+    },
+    {
+      "epoch": 0.5728924080111784,
+      "grad_norm": 0.6404868107239774,
+      "learning_rate": 4.760486794407043e-05,
+      "loss": 0.4368,
+      "step": 615
+    },
+    {
+      "epoch": 0.5775500698649279,
+      "grad_norm": 0.6133747116202044,
+      "learning_rate": 4.754013464526152e-05,
+      "loss": 0.4389,
+      "step": 620
+    },
+    {
+      "epoch": 0.5822077317186772,
+      "grad_norm": 1.0158420393440462,
+      "learning_rate": 4.747540134645262e-05,
+      "loss": 0.4492,
+      "step": 625
+    },
+    {
+      "epoch": 0.5868653935724266,
+      "grad_norm": 0.6424938062971148,
+      "learning_rate": 4.741066804764371e-05,
+      "loss": 0.4364,
+      "step": 630
+    },
+    {
+      "epoch": 0.5915230554261761,
+      "grad_norm": 0.5587540204090086,
+      "learning_rate": 4.73459347488348e-05,
+      "loss": 0.4353,
+      "step": 635
+    },
+    {
+      "epoch": 0.5961807172799255,
+      "grad_norm": 0.710771955315232,
+      "learning_rate": 4.7281201450025894e-05,
+      "loss": 0.4401,
+      "step": 640
+    },
+    {
+      "epoch": 0.6008383791336749,
+      "grad_norm": 0.8503719196899987,
+      "learning_rate": 4.721646815121699e-05,
+      "loss": 0.4401,
+      "step": 645
+    },
+    {
+      "epoch": 0.6054960409874243,
+      "grad_norm": 0.6213357524921049,
+      "learning_rate": 4.715173485240808e-05,
+      "loss": 0.4377,
+      "step": 650
+    },
+    {
+      "epoch": 0.6101537028411738,
+      "grad_norm": 0.728615015594888,
+      "learning_rate": 4.708700155359917e-05,
+      "loss": 0.4435,
+      "step": 655
+    },
+    {
+      "epoch": 0.6148113646949231,
+      "grad_norm": 0.69572702326182,
+      "learning_rate": 4.7022268254790266e-05,
+      "loss": 0.4438,
+      "step": 660
+    },
+    {
+      "epoch": 0.6194690265486725,
+      "grad_norm": 0.717769460394585,
+      "learning_rate": 4.695753495598136e-05,
+      "loss": 0.4382,
+      "step": 665
+    },
+    {
+      "epoch": 0.624126688402422,
+      "grad_norm": 0.6390733598139207,
+      "learning_rate": 4.689280165717245e-05,
+      "loss": 0.4419,
+      "step": 670
+    },
+    {
+      "epoch": 0.6287843502561714,
+      "grad_norm": 0.8175947498621025,
+      "learning_rate": 4.6828068358363545e-05,
+      "loss": 0.4338,
+      "step": 675
+    },
+    {
+      "epoch": 0.6334420121099208,
+      "grad_norm": 0.6392191276829822,
+      "learning_rate": 4.676333505955464e-05,
+      "loss": 0.437,
+      "step": 680
+    },
+    {
+      "epoch": 0.6380996739636703,
+      "grad_norm": 0.5385507245755475,
+      "learning_rate": 4.669860176074573e-05,
+      "loss": 0.448,
+      "step": 685
+    },
+    {
+      "epoch": 0.6427573358174197,
+      "grad_norm": 0.779484565903713,
+      "learning_rate": 4.6633868461936824e-05,
+      "loss": 0.4388,
+      "step": 690
+    },
+    {
+      "epoch": 0.6474149976711691,
+      "grad_norm": 0.609631979942144,
+      "learning_rate": 4.656913516312791e-05,
+      "loss": 0.4371,
+      "step": 695
+    },
+    {
+      "epoch": 0.6520726595249184,
+      "grad_norm": 0.6612175365081071,
+      "learning_rate": 4.650440186431901e-05,
+      "loss": 0.4346,
+      "step": 700
+    },
+    {
+      "epoch": 0.6567303213786679,
+      "grad_norm": 0.5776526816400351,
+      "learning_rate": 4.64396685655101e-05,
+      "loss": 0.4325,
+      "step": 705
+    },
+    {
+      "epoch": 0.6613879832324173,
+      "grad_norm": 0.6777612372991866,
+      "learning_rate": 4.637493526670119e-05,
+      "loss": 0.4433,
+      "step": 710
+    },
+    {
+      "epoch": 0.6660456450861667,
+      "grad_norm": 0.652431971316431,
+      "learning_rate": 4.631020196789229e-05,
+      "loss": 0.4383,
+      "step": 715
+    },
+    {
+      "epoch": 0.6707033069399162,
+      "grad_norm": 0.8345742527824963,
+      "learning_rate": 4.624546866908338e-05,
+      "loss": 0.4297,
+      "step": 720
+    },
+    {
+      "epoch": 0.6753609687936656,
+      "grad_norm": 0.5978893188286112,
+      "learning_rate": 4.618073537027447e-05,
+      "loss": 0.4353,
+      "step": 725
+    },
+    {
+      "epoch": 0.680018630647415,
+      "grad_norm": 0.8328268112464421,
+      "learning_rate": 4.611600207146556e-05,
+      "loss": 0.4421,
+      "step": 730
+    },
+    {
+      "epoch": 0.6846762925011645,
+      "grad_norm": 0.7087213010225971,
+      "learning_rate": 4.605126877265666e-05,
+      "loss": 0.4304,
+      "step": 735
+    },
+    {
+      "epoch": 0.6893339543549138,
+      "grad_norm": 0.6869314447013355,
+      "learning_rate": 4.598653547384775e-05,
+      "loss": 0.4347,
+      "step": 740
+    },
+    {
+      "epoch": 0.6939916162086632,
+      "grad_norm": 0.6167757431721599,
+      "learning_rate": 4.592180217503884e-05,
+      "loss": 0.4312,
+      "step": 745
+    },
+    {
+      "epoch": 0.6986492780624126,
+      "grad_norm": 0.7676543887073451,
+      "learning_rate": 4.585706887622993e-05,
+      "loss": 0.4393,
+      "step": 750
+    },
+    {
+      "epoch": 0.7033069399161621,
+      "grad_norm": 0.6961688773290436,
+      "learning_rate": 4.5792335577421026e-05,
+      "loss": 0.4295,
+      "step": 755
+    },
+    {
+      "epoch": 0.7079646017699115,
+      "grad_norm": 0.5967737066278368,
+      "learning_rate": 4.572760227861212e-05,
+      "loss": 0.4317,
+      "step": 760
+    },
+    {
+      "epoch": 0.7126222636236609,
+      "grad_norm": 0.5577548927242444,
+      "learning_rate": 4.566286897980321e-05,
+      "loss": 0.4388,
+      "step": 765
+    },
+    {
+      "epoch": 0.7172799254774104,
+      "grad_norm": 0.6798109409577441,
+      "learning_rate": 4.5598135680994305e-05,
+      "loss": 0.438,
+      "step": 770
+    },
+    {
+      "epoch": 0.7219375873311598,
+      "grad_norm": 0.7079083857791663,
+      "learning_rate": 4.55334023821854e-05,
+      "loss": 0.4266,
+      "step": 775
+    },
+    {
+      "epoch": 0.7265952491849091,
+      "grad_norm": 0.8509226139438899,
+      "learning_rate": 4.546866908337649e-05,
+      "loss": 0.4427,
+      "step": 780
+    },
+    {
+      "epoch": 0.7312529110386586,
+      "grad_norm": 0.7242979399552838,
+      "learning_rate": 4.5403935784567584e-05,
+      "loss": 0.4362,
+      "step": 785
+    },
+    {
+      "epoch": 0.735910572892408,
+      "grad_norm": 0.5877311409433356,
+      "learning_rate": 4.533920248575868e-05,
+      "loss": 0.4284,
+      "step": 790
+    },
+    {
+      "epoch": 0.7405682347461574,
+      "grad_norm": 0.6078912772108137,
+      "learning_rate": 4.527446918694977e-05,
+      "loss": 0.4347,
+      "step": 795
+    },
+    {
+      "epoch": 0.7452258965999069,
+      "grad_norm": 0.5476766413302819,
+      "learning_rate": 4.520973588814086e-05,
+      "loss": 0.4308,
+      "step": 800
+    },
+    {
+      "epoch": 0.7498835584536563,
+      "grad_norm": 0.5195378720227425,
+      "learning_rate": 4.5145002589331956e-05,
+      "loss": 0.4453,
+      "step": 805
+    },
+    {
+      "epoch": 0.7545412203074057,
+      "grad_norm": 0.7360682617560098,
+      "learning_rate": 4.508026929052305e-05,
+      "loss": 0.431,
+      "step": 810
+    },
+    {
+      "epoch": 0.759198882161155,
+      "grad_norm": 0.5685816437073101,
+      "learning_rate": 4.501553599171414e-05,
+      "loss": 0.4342,
+      "step": 815
+    },
+    {
+      "epoch": 0.7638565440149045,
+      "grad_norm": 0.5972759336495264,
+      "learning_rate": 4.495080269290523e-05,
+      "loss": 0.4336,
+      "step": 820
+    },
+    {
+      "epoch": 0.7685142058686539,
+      "grad_norm": 0.6728421551142247,
+      "learning_rate": 4.488606939409633e-05,
+      "loss": 0.4226,
+      "step": 825
+    },
+    {
+      "epoch": 0.7731718677224033,
+      "grad_norm": 0.5582332660093027,
+      "learning_rate": 4.482133609528742e-05,
+      "loss": 0.4302,
+      "step": 830
+    },
+    {
+      "epoch": 0.7778295295761528,
+      "grad_norm": 0.6980425901666201,
+      "learning_rate": 4.475660279647851e-05,
+      "loss": 0.4372,
+      "step": 835
+    },
+    {
+      "epoch": 0.7824871914299022,
+      "grad_norm": 0.7390227270424299,
+      "learning_rate": 4.46918694976696e-05,
+      "loss": 0.4232,
+      "step": 840
+    },
+    {
+      "epoch": 0.7871448532836516,
+      "grad_norm": 0.6485234571078656,
+      "learning_rate": 4.46271361988607e-05,
+      "loss": 0.4284,
+      "step": 845
+    },
+    {
+      "epoch": 0.7918025151374011,
+      "grad_norm": 0.6336992788450944,
+      "learning_rate": 4.4562402900051786e-05,
+      "loss": 0.4307,
+      "step": 850
+    },
+    {
+      "epoch": 0.7964601769911505,
+      "grad_norm": 0.7410255325128174,
+      "learning_rate": 4.449766960124288e-05,
+      "loss": 0.4236,
+      "step": 855
+    },
+    {
+      "epoch": 0.8011178388448998,
+      "grad_norm": 0.5276824982854947,
+      "learning_rate": 4.443293630243397e-05,
+      "loss": 0.4252,
+      "step": 860
+    },
+    {
+      "epoch": 0.8057755006986492,
+      "grad_norm": 0.5896389602903055,
+      "learning_rate": 4.436820300362507e-05,
+      "loss": 0.4284,
+      "step": 865
+    },
+    {
+      "epoch": 0.8104331625523987,
+      "grad_norm": 0.5253923050441579,
+      "learning_rate": 4.430346970481616e-05,
+      "loss": 0.4227,
+      "step": 870
+    },
+    {
+      "epoch": 0.8150908244061481,
+      "grad_norm": 0.5344210226388759,
+      "learning_rate": 4.423873640600725e-05,
+      "loss": 0.4321,
+      "step": 875
+    },
+    {
+      "epoch": 0.8197484862598975,
+      "grad_norm": 0.5189955942586891,
+      "learning_rate": 4.4174003107198344e-05,
+      "loss": 0.4164,
+      "step": 880
+    },
+    {
+      "epoch": 0.824406148113647,
+      "grad_norm": 0.505727185520852,
+      "learning_rate": 4.410926980838944e-05,
+      "loss": 0.4311,
+      "step": 885
+    },
+    {
+      "epoch": 0.8290638099673964,
+      "grad_norm": 0.6952374910519269,
+      "learning_rate": 4.404453650958053e-05,
+      "loss": 0.4298,
+      "step": 890
+    },
+    {
+      "epoch": 0.8337214718211458,
+      "grad_norm": 0.6334651402321975,
+      "learning_rate": 4.397980321077162e-05,
+      "loss": 0.4302,
+      "step": 895
+    },
+    {
+      "epoch": 0.8383791336748952,
+      "grad_norm": 0.5871993814882748,
+      "learning_rate": 4.3915069911962716e-05,
+      "loss": 0.4243,
+      "step": 900
+    },
+    {
+      "epoch": 0.8430367955286446,
+      "grad_norm": 0.562675211982263,
+      "learning_rate": 4.385033661315381e-05,
+      "loss": 0.4282,
+      "step": 905
+    },
+    {
+      "epoch": 0.847694457382394,
+      "grad_norm": 0.55189342404869,
+      "learning_rate": 4.37856033143449e-05,
+      "loss": 0.4342,
+      "step": 910
+    },
+    {
+      "epoch": 0.8523521192361434,
+      "grad_norm": 0.7717927793072482,
+      "learning_rate": 4.3720870015535995e-05,
+      "loss": 0.4262,
+      "step": 915
+    },
+    {
+      "epoch": 0.8570097810898929,
+      "grad_norm": 0.545706766656389,
+      "learning_rate": 4.365613671672709e-05,
+      "loss": 0.4334,
+      "step": 920
+    },
+    {
+      "epoch": 0.8616674429436423,
+      "grad_norm": 0.7308396494889845,
+      "learning_rate": 4.359140341791818e-05,
+      "loss": 0.4276,
+      "step": 925
+    },
+    {
+      "epoch": 0.8663251047973917,
+      "grad_norm": 0.6334665220388306,
+      "learning_rate": 4.352667011910927e-05,
+      "loss": 0.4289,
+      "step": 930
+    },
+    {
+      "epoch": 0.8709827666511412,
+      "grad_norm": 0.5789727565447382,
+      "learning_rate": 4.346193682030037e-05,
+      "loss": 0.4177,
+      "step": 935
+    },
+    {
+      "epoch": 0.8756404285048905,
+      "grad_norm": 0.6071364036108049,
+      "learning_rate": 4.339720352149146e-05,
+      "loss": 0.4187,
+      "step": 940
+    },
+    {
+      "epoch": 0.8802980903586399,
+      "grad_norm": 0.48355618067734446,
+      "learning_rate": 4.3332470222682546e-05,
+      "loss": 0.4222,
+      "step": 945
+    },
+    {
+      "epoch": 0.8849557522123894,
+      "grad_norm": 0.757016952941287,
+      "learning_rate": 4.326773692387364e-05,
+      "loss": 0.4149,
+      "step": 950
+    },
+    {
+      "epoch": 0.8896134140661388,
+      "grad_norm": 0.5970956354199685,
+      "learning_rate": 4.320300362506474e-05,
+      "loss": 0.4285,
+      "step": 955
+    },
+    {
+      "epoch": 0.8942710759198882,
+      "grad_norm": 1.7486796656102368,
+      "learning_rate": 4.313827032625583e-05,
+      "loss": 0.4328,
+      "step": 960
+    },
+    {
+      "epoch": 0.8989287377736377,
+      "grad_norm": 0.676896991938164,
+      "learning_rate": 4.307353702744692e-05,
+      "loss": 0.4278,
+      "step": 965
+    },
+    {
+      "epoch": 0.9035863996273871,
+      "grad_norm": 0.5343157092353157,
+      "learning_rate": 4.300880372863801e-05,
+      "loss": 0.4298,
+      "step": 970
+    },
+    {
+      "epoch": 0.9082440614811365,
+      "grad_norm": 0.6312289170160827,
+      "learning_rate": 4.294407042982911e-05,
+      "loss": 0.4237,
+      "step": 975
+    },
+    {
+      "epoch": 0.9129017233348858,
+      "grad_norm": 0.7628441977751337,
+      "learning_rate": 4.28793371310202e-05,
+      "loss": 0.4266,
+      "step": 980
+    },
+    {
+      "epoch": 0.9175593851886353,
+      "grad_norm": 1.0298339126452174,
+      "learning_rate": 4.281460383221129e-05,
+      "loss": 0.4308,
+      "step": 985
+    },
+    {
+      "epoch": 0.9222170470423847,
+      "grad_norm": 0.7772941072113776,
+      "learning_rate": 4.274987053340238e-05,
+      "loss": 0.4042,
+      "step": 990
+    },
+    {
+      "epoch": 0.9268747088961341,
+      "grad_norm": 0.8073817232990661,
+      "learning_rate": 4.2685137234593476e-05,
+      "loss": 0.4195,
+      "step": 995
+    },
+    {
+      "epoch": 0.9315323707498836,
+      "grad_norm": 0.7213209273575877,
+      "learning_rate": 4.262040393578457e-05,
+      "loss": 0.4224,
+      "step": 1000
+    },
+    {
+      "epoch": 0.936190032603633,
+      "grad_norm": 0.7416359254585871,
+      "learning_rate": 4.255567063697566e-05,
+      "loss": 0.4312,
+      "step": 1005
+    },
+    {
+      "epoch": 0.9408476944573824,
+      "grad_norm": 0.527102694728885,
+      "learning_rate": 4.2490937338166755e-05,
+      "loss": 0.4201,
+      "step": 1010
+    },
+    {
+      "epoch": 0.9455053563111319,
+      "grad_norm": 0.5594511470545082,
+      "learning_rate": 4.242620403935785e-05,
+      "loss": 0.4224,
+      "step": 1015
+    },
+    {
+      "epoch": 0.9501630181648812,
+      "grad_norm": 0.5965418648993862,
+      "learning_rate": 4.236147074054894e-05,
+      "loss": 0.4309,
+      "step": 1020
+    },
+    {
+      "epoch": 0.9548206800186306,
+      "grad_norm": 0.5450943367909993,
+      "learning_rate": 4.2296737441740034e-05,
+      "loss": 0.4227,
+      "step": 1025
+    },
+    {
+      "epoch": 0.95947834187238,
+      "grad_norm": 0.7935428149469217,
+      "learning_rate": 4.223200414293113e-05,
+      "loss": 0.422,
+      "step": 1030
+    },
+    {
+      "epoch": 0.9641360037261295,
+      "grad_norm": 0.48700743765986526,
+      "learning_rate": 4.216727084412222e-05,
+      "loss": 0.4186,
+      "step": 1035
+    },
+    {
+      "epoch": 0.9687936655798789,
+      "grad_norm": 0.7728666319795747,
+      "learning_rate": 4.2102537545313306e-05,
+      "loss": 0.4177,
+      "step": 1040
+    },
+    {
+      "epoch": 0.9734513274336283,
+      "grad_norm": 0.6153911472582105,
+      "learning_rate": 4.2037804246504406e-05,
+      "loss": 0.4185,
+      "step": 1045
+    },
+    {
+      "epoch": 0.9781089892873778,
+      "grad_norm": 0.4852026029806901,
+      "learning_rate": 4.19730709476955e-05,
+      "loss": 0.4185,
+      "step": 1050
+    },
+    {
+      "epoch": 0.9827666511411272,
+      "grad_norm": 0.4462069599855887,
+      "learning_rate": 4.190833764888659e-05,
+      "loss": 0.4193,
+      "step": 1055
+    },
+    {
+      "epoch": 0.9874243129948765,
+      "grad_norm": 0.5229879709804952,
+      "learning_rate": 4.184360435007768e-05,
+      "loss": 0.4221,
+      "step": 1060
+    },
+    {
+      "epoch": 0.992081974848626,
+      "grad_norm": 0.9581807172329343,
+      "learning_rate": 4.177887105126878e-05,
+      "loss": 0.4193,
+      "step": 1065
+    },
+    {
+      "epoch": 0.9967396367023754,
+      "grad_norm": 0.6118951656784746,
+      "learning_rate": 4.171413775245987e-05,
+      "loss": 0.411,
+      "step": 1070
+    },
+    {
+      "epoch": 1.00093153237075,
+      "grad_norm": 1.198232600503455,
+      "learning_rate": 4.164940445365096e-05,
+      "loss": 0.4222,
+      "step": 1075
+    },
+    {
+      "epoch": 1.0055891942244992,
+      "grad_norm": 0.6626140255094659,
+      "learning_rate": 4.158467115484205e-05,
+      "loss": 0.3643,
+      "step": 1080
+    },
+    {
+      "epoch": 1.0102468560782487,
+      "grad_norm": 0.6595075009077368,
+      "learning_rate": 4.151993785603315e-05,
+      "loss": 0.3533,
+      "step": 1085
+    },
+    {
+      "epoch": 1.0149045179319982,
+      "grad_norm": 0.6711382033612453,
+      "learning_rate": 4.1455204557224236e-05,
+      "loss": 0.3654,
+      "step": 1090
+    },
+    {
+      "epoch": 1.0195621797857475,
+      "grad_norm": 0.6426521900436137,
+      "learning_rate": 4.139047125841533e-05,
+      "loss": 0.3632,
+      "step": 1095
+    },
+    {
+      "epoch": 1.024219841639497,
+      "grad_norm": 0.5339378895704875,
+      "learning_rate": 4.132573795960642e-05,
+      "loss": 0.3656,
+      "step": 1100
+    },
+    {
+      "epoch": 1.0288775034932465,
+      "grad_norm": 0.5493484057627164,
+      "learning_rate": 4.1261004660797515e-05,
+      "loss": 0.3663,
+      "step": 1105
+    },
+    {
+      "epoch": 1.0335351653469957,
+      "grad_norm": 0.5273400302285899,
+      "learning_rate": 4.119627136198861e-05,
+      "loss": 0.3653,
+      "step": 1110
+    },
+    {
+      "epoch": 1.0381928272007452,
+      "grad_norm": 0.48412616047354784,
+      "learning_rate": 4.11315380631797e-05,
+      "loss": 0.3541,
+      "step": 1115
+    },
+    {
+      "epoch": 1.0428504890544947,
+      "grad_norm": 0.6089370186461038,
+      "learning_rate": 4.1066804764370794e-05,
+      "loss": 0.3635,
+      "step": 1120
+    },
+    {
+      "epoch": 1.047508150908244,
+      "grad_norm": 0.6152489330197201,
+      "learning_rate": 4.100207146556189e-05,
+      "loss": 0.3692,
+      "step": 1125
+    },
+    {
+      "epoch": 1.0521658127619935,
+      "grad_norm": 0.4903458431092099,
+      "learning_rate": 4.093733816675298e-05,
+      "loss": 0.3528,
+      "step": 1130
+    },
+    {
+      "epoch": 1.056823474615743,
+      "grad_norm": 0.45218893931929327,
+      "learning_rate": 4.087260486794407e-05,
+      "loss": 0.3641,
+      "step": 1135
+    },
+    {
+      "epoch": 1.0614811364694923,
+      "grad_norm": 0.5634215362291429,
+      "learning_rate": 4.0807871569135166e-05,
+      "loss": 0.3633,
+      "step": 1140
+    },
+    {
+      "epoch": 1.0661387983232418,
+      "grad_norm": 0.44840145346738985,
+      "learning_rate": 4.074313827032626e-05,
+      "loss": 0.3634,
+      "step": 1145
+    },
+    {
+      "epoch": 1.0707964601769913,
+      "grad_norm": 0.47176553149630446,
+      "learning_rate": 4.067840497151735e-05,
+      "loss": 0.3586,
+      "step": 1150
+    },
+    {
+      "epoch": 1.0754541220307405,
+      "grad_norm": 0.5804511973024695,
+      "learning_rate": 4.0613671672708445e-05,
+      "loss": 0.3622,
+      "step": 1155
+    },
+    {
+      "epoch": 1.08011178388449,
+      "grad_norm": 0.5594430865141662,
+      "learning_rate": 4.054893837389954e-05,
+      "loss": 0.3699,
+      "step": 1160
+    },
+    {
+      "epoch": 1.0847694457382393,
+      "grad_norm": 0.4058236118445243,
+      "learning_rate": 4.048420507509063e-05,
+      "loss": 0.3705,
+      "step": 1165
+    },
+    {
+      "epoch": 1.0894271075919888,
+      "grad_norm": 0.5048270986690303,
+      "learning_rate": 4.041947177628172e-05,
+      "loss": 0.3675,
+      "step": 1170
+    },
+    {
+      "epoch": 1.0940847694457383,
+      "grad_norm": 0.5363220404825172,
+      "learning_rate": 4.035473847747282e-05,
+      "loss": 0.3684,
+      "step": 1175
+    },
+    {
+      "epoch": 1.0987424312994876,
+      "grad_norm": 0.5199102851213286,
+      "learning_rate": 4.029000517866391e-05,
+      "loss": 0.36,
+      "step": 1180
+    },
+    {
+      "epoch": 1.103400093153237,
+      "grad_norm": 0.5335401206236534,
+      "learning_rate": 4.0225271879854996e-05,
+      "loss": 0.3611,
+      "step": 1185
+    },
+    {
+      "epoch": 1.1080577550069866,
+      "grad_norm": 0.3877188665567559,
+      "learning_rate": 4.016053858104609e-05,
+      "loss": 0.3428,
+      "step": 1190
+    },
+    {
+      "epoch": 1.1127154168607358,
+      "grad_norm": 0.47198305564893206,
+      "learning_rate": 4.009580528223719e-05,
+      "loss": 0.3521,
+      "step": 1195
+    },
+    {
+      "epoch": 1.1173730787144853,
+      "grad_norm": 0.4375832663471869,
+      "learning_rate": 4.0031071983428275e-05,
+      "loss": 0.36,
+      "step": 1200
+    },
+    {
+      "epoch": 1.1220307405682348,
+      "grad_norm": 0.5734782102745869,
+      "learning_rate": 3.996633868461937e-05,
+      "loss": 0.3692,
+      "step": 1205
+    },
+    {
+      "epoch": 1.126688402421984,
+      "grad_norm": 0.466632874743288,
+      "learning_rate": 3.990160538581046e-05,
+      "loss": 0.3629,
+      "step": 1210
+    },
+    {
+      "epoch": 1.1313460642757336,
+      "grad_norm": 0.6219265879975726,
+      "learning_rate": 3.9836872087001554e-05,
+      "loss": 0.3595,
+      "step": 1215
+    },
+    {
+      "epoch": 1.136003726129483,
+      "grad_norm": 0.6585230588680556,
+      "learning_rate": 3.977213878819265e-05,
+      "loss": 0.3759,
+      "step": 1220
+    },
+    {
+      "epoch": 1.1406613879832324,
+      "grad_norm": 0.625960144911052,
+      "learning_rate": 3.970740548938374e-05,
+      "loss": 0.3608,
+      "step": 1225
+    },
+    {
+      "epoch": 1.1453190498369819,
+      "grad_norm": 0.5613416054644293,
+      "learning_rate": 3.964267219057483e-05,
+      "loss": 0.3591,
+      "step": 1230
+    },
+    {
+      "epoch": 1.1499767116907313,
+      "grad_norm": 0.6293922717444177,
+      "learning_rate": 3.9577938891765926e-05,
+      "loss": 0.3633,
+      "step": 1235
+    },
+    {
+      "epoch": 1.1546343735444806,
+      "grad_norm": 0.4382344516265855,
+      "learning_rate": 3.951320559295702e-05,
+      "loss": 0.3673,
+      "step": 1240
+    },
+    {
+      "epoch": 1.1592920353982301,
+      "grad_norm": 0.43635813377518307,
+      "learning_rate": 3.944847229414811e-05,
+      "loss": 0.3694,
+      "step": 1245
+    },
+    {
+      "epoch": 1.1639496972519794,
+      "grad_norm": 0.6486802446293214,
+      "learning_rate": 3.9383738995339205e-05,
+      "loss": 0.3721,
+      "step": 1250
+    },
+    {
+      "epoch": 1.1686073591057289,
+      "grad_norm": 0.5468009019040315,
+      "learning_rate": 3.93190056965303e-05,
+      "loss": 0.3594,
+      "step": 1255
+    },
+    {
+      "epoch": 1.1732650209594784,
+      "grad_norm": 0.4992784524933434,
+      "learning_rate": 3.925427239772139e-05,
+      "loss": 0.3723,
+      "step": 1260
+    },
+    {
+      "epoch": 1.1779226828132279,
+      "grad_norm": 0.48099901440871706,
+      "learning_rate": 3.9189539098912484e-05,
+      "loss": 0.3733,
+      "step": 1265
+    },
+    {
+      "epoch": 1.1825803446669771,
+      "grad_norm": 0.5099848543509489,
+      "learning_rate": 3.912480580010358e-05,
+      "loss": 0.3531,
+      "step": 1270
+    },
+    {
+      "epoch": 1.1872380065207266,
+      "grad_norm": 0.5202936478101648,
+      "learning_rate": 3.906007250129467e-05,
+      "loss": 0.3637,
+      "step": 1275
+    },
+    {
+      "epoch": 1.191895668374476,
+      "grad_norm": 0.534882109326889,
+      "learning_rate": 3.8995339202485756e-05,
+      "loss": 0.3689,
+      "step": 1280
+    },
+    {
+      "epoch": 1.1965533302282254,
+      "grad_norm": 0.5919980159275785,
+      "learning_rate": 3.8930605903676856e-05,
+      "loss": 0.3712,
+      "step": 1285
+    },
+    {
+      "epoch": 1.201210992081975,
+      "grad_norm": 0.6393674652244553,
+      "learning_rate": 3.886587260486795e-05,
+      "loss": 0.3609,
+      "step": 1290
+    },
+    {
+      "epoch": 1.2058686539357242,
+      "grad_norm": 0.449384764222771,
+      "learning_rate": 3.8801139306059035e-05,
+      "loss": 0.3633,
+      "step": 1295
+    },
+    {
+      "epoch": 1.2105263157894737,
+      "grad_norm": 0.5892594923823059,
+      "learning_rate": 3.873640600725013e-05,
+      "loss": 0.3595,
+      "step": 1300
+    },
+    {
+      "epoch": 1.2151839776432232,
+      "grad_norm": 0.5334687166327565,
+      "learning_rate": 3.867167270844123e-05,
+      "loss": 0.3699,
+      "step": 1305
+    },
+    {
+      "epoch": 1.2198416394969724,
+      "grad_norm": 0.44268921628250013,
+      "learning_rate": 3.8606939409632314e-05,
+      "loss": 0.3722,
+      "step": 1310
+    },
+    {
+      "epoch": 1.224499301350722,
+      "grad_norm": 0.4291582971347598,
+      "learning_rate": 3.854220611082341e-05,
+      "loss": 0.3619,
+      "step": 1315
+    },
+    {
+      "epoch": 1.2291569632044714,
+      "grad_norm": 0.4336800197463241,
+      "learning_rate": 3.84774728120145e-05,
+      "loss": 0.3715,
+      "step": 1320
+    },
+    {
+      "epoch": 1.2338146250582207,
+      "grad_norm": 0.4443708683400446,
+      "learning_rate": 3.841273951320559e-05,
+      "loss": 0.3622,
+      "step": 1325
+    },
+    {
+      "epoch": 1.2384722869119702,
+      "grad_norm": 0.427129846430766,
+      "learning_rate": 3.8348006214396686e-05,
+      "loss": 0.3692,
+      "step": 1330
+    },
+    {
+      "epoch": 1.2431299487657197,
+      "grad_norm": 0.5591150556223793,
+      "learning_rate": 3.828327291558778e-05,
+      "loss": 0.3622,
+      "step": 1335
+    },
+    {
+      "epoch": 1.247787610619469,
+      "grad_norm": 0.5657209049937582,
+      "learning_rate": 3.821853961677888e-05,
+      "loss": 0.3637,
+      "step": 1340
+    },
+    {
+      "epoch": 1.2524452724732185,
+      "grad_norm": 0.49877057776070494,
+      "learning_rate": 3.8153806317969965e-05,
+      "loss": 0.3533,
+      "step": 1345
+    },
+    {
+      "epoch": 1.257102934326968,
+      "grad_norm": 0.5817369057575189,
+      "learning_rate": 3.808907301916106e-05,
+      "loss": 0.3633,
+      "step": 1350
+    },
+    {
+      "epoch": 1.2617605961807172,
+      "grad_norm": 0.5874812736984852,
+      "learning_rate": 3.802433972035215e-05,
+      "loss": 0.3594,
+      "step": 1355
+    },
+    {
+      "epoch": 1.2664182580344667,
+      "grad_norm": 0.6207126846666516,
+      "learning_rate": 3.7959606421543244e-05,
+      "loss": 0.3528,
+      "step": 1360
+    },
+    {
+      "epoch": 1.271075919888216,
+      "grad_norm": 0.5366782936079324,
+      "learning_rate": 3.789487312273434e-05,
+      "loss": 0.3656,
+      "step": 1365
+    },
+    {
+      "epoch": 1.2757335817419655,
+      "grad_norm": 0.5830667256592843,
+      "learning_rate": 3.783013982392543e-05,
+      "loss": 0.3634,
+      "step": 1370
+    },
+    {
+      "epoch": 1.280391243595715,
+      "grad_norm": 0.535133020380144,
+      "learning_rate": 3.776540652511652e-05,
+      "loss": 0.3638,
+      "step": 1375
+    },
+    {
+      "epoch": 1.2850489054494645,
+      "grad_norm": 0.5605033732100219,
+      "learning_rate": 3.7700673226307616e-05,
+      "loss": 0.3634,
+      "step": 1380
+    },
+    {
+      "epoch": 1.2897065673032138,
+      "grad_norm": 0.6454282897245199,
+      "learning_rate": 3.763593992749871e-05,
+      "loss": 0.3617,
+      "step": 1385
+    },
+    {
+      "epoch": 1.2943642291569633,
+      "grad_norm": 0.569922823326428,
+      "learning_rate": 3.7571206628689795e-05,
+      "loss": 0.3715,
+      "step": 1390
+    },
+    {
+      "epoch": 1.2990218910107125,
+      "grad_norm": 0.7489830344450082,
+      "learning_rate": 3.7506473329880895e-05,
+      "loss": 0.4492,
+      "step": 1395
+    },
+    {
+      "epoch": 1.303679552864462,
+      "grad_norm": 0.5804151095364183,
+      "learning_rate": 3.744174003107199e-05,
+      "loss": 0.3672,
+      "step": 1400
+    },
+    {
+      "epoch": 1.3083372147182115,
+      "grad_norm": 0.45978003988666893,
+      "learning_rate": 3.7377006732263074e-05,
+      "loss": 0.3602,
+      "step": 1405
+    },
+    {
+      "epoch": 1.312994876571961,
+      "grad_norm": 0.4805974767093187,
+      "learning_rate": 3.731227343345417e-05,
+      "loss": 0.3589,
+      "step": 1410
+    },
+    {
+      "epoch": 1.3176525384257103,
+      "grad_norm": 0.4846107371473044,
+      "learning_rate": 3.724754013464527e-05,
+      "loss": 0.3775,
+      "step": 1415
+    },
+    {
+      "epoch": 1.3223102002794598,
+      "grad_norm": 0.44140763746654454,
+      "learning_rate": 3.718280683583635e-05,
+      "loss": 0.3649,
+      "step": 1420
+    },
+    {
+      "epoch": 1.326967862133209,
+      "grad_norm": 0.3981869098899127,
+      "learning_rate": 3.7118073537027446e-05,
+      "loss": 0.3609,
+      "step": 1425
+    },
+    {
+      "epoch": 1.3316255239869585,
+      "grad_norm": 0.6646189023622765,
+      "learning_rate": 3.705334023821854e-05,
+      "loss": 0.361,
+      "step": 1430
+    },
+    {
+      "epoch": 1.336283185840708,
+      "grad_norm": 0.5190775947864457,
+      "learning_rate": 3.698860693940964e-05,
+      "loss": 0.3651,
+      "step": 1435
+    },
+    {
+      "epoch": 1.3409408476944573,
+      "grad_norm": 0.4665782605567536,
+      "learning_rate": 3.6923873640600725e-05,
+      "loss": 0.3637,
+      "step": 1440
+    },
+    {
+      "epoch": 1.3455985095482068,
+      "grad_norm": 0.533120751391509,
+      "learning_rate": 3.685914034179182e-05,
+      "loss": 0.3705,
+      "step": 1445
+    },
+    {
+      "epoch": 1.350256171401956,
+      "grad_norm": 0.4454321342916118,
+      "learning_rate": 3.679440704298292e-05,
+      "loss": 0.3606,
+      "step": 1450
+    },
+    {
+      "epoch": 1.3549138332557056,
+      "grad_norm": 0.43930714074978666,
+      "learning_rate": 3.6729673744174004e-05,
+      "loss": 0.3699,
+      "step": 1455
+    },
+    {
+      "epoch": 1.359571495109455,
+      "grad_norm": 0.5051073774169964,
+      "learning_rate": 3.66649404453651e-05,
+      "loss": 0.3542,
+      "step": 1460
+    },
+    {
+      "epoch": 1.3642291569632046,
+      "grad_norm": 0.4914786128754226,
+      "learning_rate": 3.660020714655619e-05,
+      "loss": 0.3611,
+      "step": 1465
+    },
+    {
+      "epoch": 1.3688868188169538,
+      "grad_norm": 0.48917577891457625,
+      "learning_rate": 3.653547384774728e-05,
+      "loss": 0.3647,
+      "step": 1470
+    },
+    {
+      "epoch": 1.3735444806707033,
+      "grad_norm": 0.5218401916805995,
+      "learning_rate": 3.6470740548938376e-05,
+      "loss": 0.3601,
+      "step": 1475
+    },
+    {
+      "epoch": 1.3782021425244526,
+      "grad_norm": 0.4542309075740912,
+      "learning_rate": 3.640600725012947e-05,
+      "loss": 0.3603,
+      "step": 1480
+    },
+    {
+      "epoch": 1.382859804378202,
+      "grad_norm": 0.593114065460838,
+      "learning_rate": 3.634127395132056e-05,
+      "loss": 0.3667,
+      "step": 1485
+    },
+    {
+      "epoch": 1.3875174662319516,
+      "grad_norm": 0.7061630570341471,
+      "learning_rate": 3.6276540652511655e-05,
+      "loss": 0.3684,
+      "step": 1490
+    },
+    {
+      "epoch": 1.392175128085701,
+      "grad_norm": 1.1149409130136005,
+      "learning_rate": 3.621180735370275e-05,
+      "loss": 0.4374,
+      "step": 1495
+    },
+    {
+      "epoch": 1.3968327899394504,
+      "grad_norm": 1.7493220952744504,
+      "learning_rate": 3.6147074054893834e-05,
+      "loss": 0.7659,
+      "step": 1500
+    },
+    {
+      "epoch": 1.4014904517931999,
+      "grad_norm": 0.8169838033128435,
+      "learning_rate": 3.6082340756084934e-05,
+      "loss": 0.3708,
+      "step": 1505
+    },
+    {
+      "epoch": 1.4061481136469491,
+      "grad_norm": 0.6343950966738808,
+      "learning_rate": 3.601760745727603e-05,
+      "loss": 0.3651,
+      "step": 1510
+    },
+    {
+      "epoch": 1.4108057755006986,
+      "grad_norm": 0.722357376224152,
+      "learning_rate": 3.595287415846711e-05,
+      "loss": 0.3735,
+      "step": 1515
+    },
+    {
+      "epoch": 1.4154634373544481,
+      "grad_norm": 0.5665404433376358,
+      "learning_rate": 3.5888140859658206e-05,
+      "loss": 0.3658,
+      "step": 1520
+    },
+    {
+      "epoch": 1.4201210992081974,
+      "grad_norm": 0.5893727155297599,
+      "learning_rate": 3.5823407560849306e-05,
+      "loss": 0.3653,
+      "step": 1525
+    },
+    {
+      "epoch": 1.424778761061947,
+      "grad_norm": 0.5476427446105051,
+      "learning_rate": 3.57586742620404e-05,
+      "loss": 0.3557,
+      "step": 1530
+    },
+    {
+      "epoch": 1.4294364229156964,
+      "grad_norm": 0.691182790002671,
+      "learning_rate": 3.5693940963231485e-05,
+      "loss": 0.3706,
+      "step": 1535
+    },
+    {
+      "epoch": 1.4340940847694457,
+      "grad_norm": 0.46156409326817516,
+      "learning_rate": 3.5629207664422585e-05,
+      "loss": 0.3614,
+      "step": 1540
+    },
+    {
+      "epoch": 1.4387517466231952,
+      "grad_norm": 0.4598532920012943,
+      "learning_rate": 3.556447436561368e-05,
+      "loss": 0.3631,
+      "step": 1545
+    },
+    {
+      "epoch": 1.4434094084769447,
+      "grad_norm": 0.4380166097581896,
+      "learning_rate": 3.5499741066804764e-05,
+      "loss": 0.3583,
+      "step": 1550
+    },
+    {
+      "epoch": 1.448067070330694,
+      "grad_norm": 0.4064716472176242,
+      "learning_rate": 3.543500776799586e-05,
+      "loss": 0.3533,
+      "step": 1555
+    },
+    {
+      "epoch": 1.4527247321844434,
+      "grad_norm": 0.4548851529702526,
+      "learning_rate": 3.537027446918696e-05,
+      "loss": 0.3599,
+      "step": 1560
+    },
+    {
+      "epoch": 1.4573823940381927,
+      "grad_norm": 0.36916212723651665,
+      "learning_rate": 3.530554117037804e-05,
+      "loss": 0.3685,
+      "step": 1565
+    },
+    {
+      "epoch": 1.4620400558919422,
+      "grad_norm": 0.5066725999452286,
+      "learning_rate": 3.5240807871569136e-05,
+      "loss": 0.381,
+      "step": 1570
+    },
+    {
+      "epoch": 1.4666977177456917,
+      "grad_norm": 0.4774024861164161,
+      "learning_rate": 3.517607457276023e-05,
+      "loss": 0.3635,
+      "step": 1575
+    },
+    {
+      "epoch": 1.4713553795994412,
+      "grad_norm": 0.5930294981464446,
+      "learning_rate": 3.511134127395132e-05,
+      "loss": 0.355,
+      "step": 1580
+    },
+    {
+      "epoch": 1.4760130414531905,
+      "grad_norm": 0.6683615507217858,
+      "learning_rate": 3.5046607975142415e-05,
+      "loss": 0.3632,
+      "step": 1585
+    },
+    {
+      "epoch": 1.48067070330694,
+      "grad_norm": 0.47781095190925266,
+      "learning_rate": 3.498187467633351e-05,
+      "loss": 0.3651,
+      "step": 1590
+    },
+    {
+      "epoch": 1.4853283651606892,
+      "grad_norm": 0.4816343305435411,
+      "learning_rate": 3.49171413775246e-05,
+      "loss": 0.361,
+      "step": 1595
+    },
+    {
+      "epoch": 1.4899860270144387,
+      "grad_norm": 0.36945600164162296,
+      "learning_rate": 3.4852408078715694e-05,
+      "loss": 0.3664,
+      "step": 1600
+    },
+    {
+      "epoch": 1.4946436888681882,
+      "grad_norm": 0.42161998754880686,
+      "learning_rate": 3.478767477990679e-05,
+      "loss": 0.3566,
+      "step": 1605
+    },
+    {
+      "epoch": 1.4993013507219377,
+      "grad_norm": 0.43171270393570554,
+      "learning_rate": 3.472294148109787e-05,
+      "loss": 0.3638,
+      "step": 1610
+    },
+    {
+      "epoch": 1.503959012575687,
+      "grad_norm": 0.49449118003805265,
+      "learning_rate": 3.465820818228897e-05,
+      "loss": 0.3616,
+      "step": 1615
+    },
+    {
+      "epoch": 1.5086166744294365,
+      "grad_norm": 0.5126145855254146,
+      "learning_rate": 3.4593474883480066e-05,
+      "loss": 0.3627,
+      "step": 1620
+    },
+    {
+      "epoch": 1.5132743362831858,
+      "grad_norm": 0.4825580222932842,
+      "learning_rate": 3.452874158467116e-05,
+      "loss": 0.3666,
+      "step": 1625
+    },
+    {
+      "epoch": 1.5179319981369352,
+      "grad_norm": 0.5265848281578551,
+      "learning_rate": 3.4464008285862245e-05,
+      "loss": 0.3633,
+      "step": 1630
+    },
+    {
+      "epoch": 1.5225896599906847,
+      "grad_norm": 0.4810011855595015,
+      "learning_rate": 3.4399274987053345e-05,
+      "loss": 0.3748,
+      "step": 1635
+    },
+    {
+      "epoch": 1.5272473218444342,
+      "grad_norm": 0.4566927190333215,
+      "learning_rate": 3.433454168824444e-05,
+      "loss": 0.3673,
+      "step": 1640
+    },
+    {
+      "epoch": 1.5319049836981835,
+      "grad_norm": 0.46400710872493467,
+      "learning_rate": 3.4269808389435524e-05,
+      "loss": 0.3711,
+      "step": 1645
+    },
+    {
+      "epoch": 1.5365626455519328,
+      "grad_norm": 3.1087721003712434,
+      "learning_rate": 3.4205075090626624e-05,
+      "loss": 0.356,
+      "step": 1650
+    },
+    {
+      "epoch": 1.5412203074056823,
+      "grad_norm": 0.5034579559440693,
+      "learning_rate": 3.414034179181772e-05,
+      "loss": 0.3552,
+      "step": 1655
+    },
+    {
+      "epoch": 1.5458779692594318,
+      "grad_norm": 0.4780598324542499,
+      "learning_rate": 3.40756084930088e-05,
+      "loss": 0.355,
+      "step": 1660
+    },
+    {
+      "epoch": 1.5505356311131813,
+      "grad_norm": 0.5764212486237088,
+      "learning_rate": 3.4010875194199896e-05,
+      "loss": 0.3676,
+      "step": 1665
+    },
+    {
+      "epoch": 1.5551932929669308,
+      "grad_norm": 0.5517039007987018,
+      "learning_rate": 3.3946141895390996e-05,
+      "loss": 0.3636,
+      "step": 1670
+    },
+    {
+      "epoch": 1.55985095482068,
+      "grad_norm": 0.5058857939828647,
+      "learning_rate": 3.388140859658208e-05,
+      "loss": 0.362,
+      "step": 1675
+    },
+    {
+      "epoch": 1.5645086166744293,
+      "grad_norm": 0.43254113359686863,
+      "learning_rate": 3.3816675297773175e-05,
+      "loss": 0.3689,
+      "step": 1680
+    },
+    {
+      "epoch": 1.5691662785281788,
+      "grad_norm": 0.4639020956548089,
+      "learning_rate": 3.375194199896427e-05,
+      "loss": 0.3571,
+      "step": 1685
+    },
+    {
+      "epoch": 1.5738239403819283,
+      "grad_norm": 0.5001606568137839,
+      "learning_rate": 3.368720870015536e-05,
+      "loss": 0.3644,
+      "step": 1690
+    },
+    {
+      "epoch": 1.5784816022356778,
+      "grad_norm": 0.4373315768297557,
+      "learning_rate": 3.3622475401346454e-05,
+      "loss": 0.3612,
+      "step": 1695
+    },
+    {
+      "epoch": 1.583139264089427,
+      "grad_norm": 0.4139251174179008,
+      "learning_rate": 3.355774210253755e-05,
+      "loss": 0.3658,
+      "step": 1700
+    },
+    {
+      "epoch": 1.5877969259431766,
+      "grad_norm": 0.4322065090656477,
+      "learning_rate": 3.349300880372864e-05,
+      "loss": 0.3591,
+      "step": 1705
+    },
+    {
+      "epoch": 1.5924545877969258,
+      "grad_norm": 0.4970626082566968,
+      "learning_rate": 3.342827550491973e-05,
+      "loss": 0.3663,
+      "step": 1710
+    },
+    {
+      "epoch": 1.5971122496506753,
+      "grad_norm": 0.4705280367004464,
+      "learning_rate": 3.3363542206110826e-05,
+      "loss": 0.3671,
+      "step": 1715
+    },
+    {
+      "epoch": 1.6017699115044248,
+      "grad_norm": 0.43124848527955606,
+      "learning_rate": 3.329880890730191e-05,
+      "loss": 0.3537,
+      "step": 1720
+    },
+    {
+      "epoch": 1.6064275733581743,
+      "grad_norm": 0.5258412470469775,
+      "learning_rate": 3.323407560849301e-05,
+      "loss": 0.3603,
+      "step": 1725
+    },
+    {
+      "epoch": 1.6110852352119236,
+      "grad_norm": 0.4396848395121731,
+      "learning_rate": 3.3169342309684105e-05,
+      "loss": 0.372,
+      "step": 1730
+    },
+    {
+      "epoch": 1.6157428970656729,
+      "grad_norm": 0.4275239821085271,
+      "learning_rate": 3.31046090108752e-05,
+      "loss": 0.3715,
+      "step": 1735
+    },
+    {
+      "epoch": 1.6204005589194224,
+      "grad_norm": 0.47043887246154587,
+      "learning_rate": 3.3039875712066284e-05,
+      "loss": 0.3523,
+      "step": 1740
+    },
+    {
+      "epoch": 1.6250582207731719,
+      "grad_norm": 0.4830833939098779,
+      "learning_rate": 3.2975142413257384e-05,
+      "loss": 0.3823,
+      "step": 1745
+    },
+    {
+      "epoch": 1.6297158826269214,
+      "grad_norm": 0.363713578471795,
+      "learning_rate": 3.291040911444848e-05,
+      "loss": 0.3601,
+      "step": 1750
+    },
+    {
+      "epoch": 1.6343735444806708,
+      "grad_norm": 0.4199388157747959,
+      "learning_rate": 3.284567581563956e-05,
+      "loss": 0.3702,
+      "step": 1755
+    },
+    {
+      "epoch": 1.6390312063344201,
+      "grad_norm": 0.42626426655806804,
+      "learning_rate": 3.278094251683066e-05,
+      "loss": 0.3554,
+      "step": 1760
+    },
+    {
+      "epoch": 1.6436888681881694,
+      "grad_norm": 0.567374135835316,
+      "learning_rate": 3.2716209218021756e-05,
+      "loss": 0.359,
+      "step": 1765
+    },
+    {
+      "epoch": 1.648346530041919,
+      "grad_norm": 0.43110033232228706,
+      "learning_rate": 3.265147591921284e-05,
+      "loss": 0.3542,
+      "step": 1770
+    },
+    {
+      "epoch": 1.6530041918956684,
+      "grad_norm": 0.5238869056328617,
+      "learning_rate": 3.2586742620403935e-05,
+      "loss": 0.3678,
+      "step": 1775
+    },
+    {
+      "epoch": 1.6576618537494179,
+      "grad_norm": 0.49792087625873993,
+      "learning_rate": 3.2522009321595035e-05,
+      "loss": 0.3566,
+      "step": 1780
+    },
+    {
+      "epoch": 1.6623195156031674,
+      "grad_norm": 0.44022742289658967,
+      "learning_rate": 3.245727602278612e-05,
+      "loss": 0.3644,
+      "step": 1785
+    },
+    {
+      "epoch": 1.6669771774569166,
+      "grad_norm": 0.4020789247285759,
+      "learning_rate": 3.2392542723977214e-05,
+      "loss": 0.3558,
+      "step": 1790
+    },
+    {
+      "epoch": 1.671634839310666,
+      "grad_norm": 0.46941803599205045,
+      "learning_rate": 3.232780942516831e-05,
+      "loss": 0.3554,
+      "step": 1795
+    },
+    {
+      "epoch": 1.6762925011644154,
+      "grad_norm": 0.44615738959074275,
+      "learning_rate": 3.22630761263594e-05,
+      "loss": 0.3532,
+      "step": 1800
+    },
+    {
+      "epoch": 1.680950163018165,
+      "grad_norm": 0.5307510022659958,
+      "learning_rate": 3.219834282755049e-05,
+      "loss": 0.3613,
+      "step": 1805
+    },
+    {
+      "epoch": 1.6856078248719144,
+      "grad_norm": 0.43074684583169204,
+      "learning_rate": 3.2133609528741586e-05,
+      "loss": 0.3596,
+      "step": 1810
+    },
+    {
+      "epoch": 1.6902654867256637,
+      "grad_norm": 0.45453772826067634,
+      "learning_rate": 3.206887622993268e-05,
+      "loss": 0.3605,
+      "step": 1815
+    },
+    {
+      "epoch": 1.6949231485794132,
+      "grad_norm": 0.39694233100152826,
+      "learning_rate": 3.200414293112377e-05,
+      "loss": 0.3484,
+      "step": 1820
+    },
+    {
+      "epoch": 1.6995808104331624,
+      "grad_norm": 0.48125790946047087,
+      "learning_rate": 3.1939409632314865e-05,
+      "loss": 0.351,
+      "step": 1825
+    },
+    {
+      "epoch": 1.704238472286912,
+      "grad_norm": 0.42999870591028383,
+      "learning_rate": 3.187467633350596e-05,
+      "loss": 0.356,
+      "step": 1830
+    },
+    {
+      "epoch": 1.7088961341406614,
+      "grad_norm": 0.46596362181549944,
+      "learning_rate": 3.180994303469705e-05,
+      "loss": 0.3632,
+      "step": 1835
+    },
+    {
+      "epoch": 1.713553795994411,
+      "grad_norm": 0.4325609693374574,
+      "learning_rate": 3.1745209735888144e-05,
+      "loss": 0.3575,
+      "step": 1840
+    },
+    {
+      "epoch": 1.7182114578481602,
+      "grad_norm": 0.41540892625043885,
+      "learning_rate": 3.168047643707924e-05,
+      "loss": 0.3628,
+      "step": 1845
+    },
+    {
+      "epoch": 1.7228691197019095,
+      "grad_norm": 0.4031078505421204,
+      "learning_rate": 3.161574313827032e-05,
+      "loss": 0.3703,
+      "step": 1850
+    },
+    {
+      "epoch": 1.727526781555659,
+      "grad_norm": 0.41838515198339477,
+      "learning_rate": 3.155100983946142e-05,
+      "loss": 0.3518,
+      "step": 1855
+    },
+    {
+      "epoch": 1.7321844434094085,
+      "grad_norm": 0.45646943466714157,
+      "learning_rate": 3.1486276540652516e-05,
+      "loss": 0.3755,
+      "step": 1860
+    },
+    {
+      "epoch": 1.736842105263158,
+      "grad_norm": 0.585084433635785,
+      "learning_rate": 3.14215432418436e-05,
+      "loss": 0.3552,
+      "step": 1865
+    },
+    {
+      "epoch": 1.7414997671169075,
+      "grad_norm": 0.3911867696777629,
+      "learning_rate": 3.13568099430347e-05,
+      "loss": 0.3595,
+      "step": 1870
+    },
+    {
+      "epoch": 1.7461574289706567,
+      "grad_norm": 0.4680970227691401,
+      "learning_rate": 3.1292076644225795e-05,
+      "loss": 0.3749,
+      "step": 1875
+    },
+    {
+      "epoch": 1.750815090824406,
+      "grad_norm": 0.5489571566456214,
+      "learning_rate": 3.122734334541688e-05,
+      "loss": 0.3652,
+      "step": 1880
+    },
+    {
+      "epoch": 1.7554727526781555,
+      "grad_norm": 0.539830040482289,
+      "learning_rate": 3.1162610046607974e-05,
+      "loss": 0.3672,
+      "step": 1885
+    },
+    {
+      "epoch": 1.760130414531905,
+      "grad_norm": 0.540754916595141,
+      "learning_rate": 3.1097876747799074e-05,
+      "loss": 0.3641,
+      "step": 1890
+    },
+    {
+      "epoch": 1.7647880763856545,
+      "grad_norm": 0.5219699672209992,
+      "learning_rate": 3.103314344899016e-05,
+      "loss": 0.373,
+      "step": 1895
+    },
+    {
+      "epoch": 1.7694457382394038,
+      "grad_norm": 0.377337385788726,
+      "learning_rate": 3.096841015018125e-05,
+      "loss": 0.3611,
+      "step": 1900
+    },
+    {
+      "epoch": 1.7741034000931533,
+      "grad_norm": 0.48563812072688267,
+      "learning_rate": 3.0903676851372346e-05,
+      "loss": 0.3625,
+      "step": 1905
+    },
+    {
+      "epoch": 1.7787610619469025,
+      "grad_norm": 0.5845893805457868,
+      "learning_rate": 3.083894355256344e-05,
+      "loss": 0.3551,
+      "step": 1910
+    },
+    {
+      "epoch": 1.783418723800652,
+      "grad_norm": 0.5664348577055764,
+      "learning_rate": 3.077421025375453e-05,
+      "loss": 0.3508,
+      "step": 1915
+    },
+    {
+      "epoch": 1.7880763856544015,
+      "grad_norm": 0.4921575417307053,
+      "learning_rate": 3.0709476954945625e-05,
+      "loss": 0.3564,
+      "step": 1920
+    },
+    {
+      "epoch": 1.792734047508151,
+      "grad_norm": 0.41530413831923074,
+      "learning_rate": 3.064474365613672e-05,
+      "loss": 0.3644,
+      "step": 1925
+    },
+    {
+      "epoch": 1.7973917093619003,
+      "grad_norm": 0.45996940636819483,
+      "learning_rate": 3.058001035732781e-05,
+      "loss": 0.354,
+      "step": 1930
+    },
+    {
+      "epoch": 1.8020493712156498,
+      "grad_norm": 0.5159015527502255,
+      "learning_rate": 3.0515277058518904e-05,
+      "loss": 0.3509,
+      "step": 1935
+    },
+    {
+      "epoch": 1.806707033069399,
+      "grad_norm": 0.5152021259341406,
+      "learning_rate": 3.045054375971e-05,
+      "loss": 0.353,
+      "step": 1940
+    },
+    {
+      "epoch": 1.8113646949231486,
+      "grad_norm": 0.42063698743452166,
+      "learning_rate": 3.0385810460901086e-05,
+      "loss": 0.3635,
+      "step": 1945
+    },
+    {
+      "epoch": 1.816022356776898,
+      "grad_norm": 0.4197227843543687,
+      "learning_rate": 3.0321077162092183e-05,
+      "loss": 0.3552,
+      "step": 1950
+    },
+    {
+      "epoch": 1.8206800186306475,
+      "grad_norm": 0.4386043250783782,
+      "learning_rate": 3.0256343863283276e-05,
+      "loss": 0.3584,
+      "step": 1955
+    },
+    {
+      "epoch": 1.8253376804843968,
+      "grad_norm": 0.43078152623180066,
+      "learning_rate": 3.0191610564474365e-05,
+      "loss": 0.3594,
+      "step": 1960
+    },
+    {
+      "epoch": 1.829995342338146,
+      "grad_norm": 0.41453786330622155,
+      "learning_rate": 3.0126877265665458e-05,
+      "loss": 0.3617,
+      "step": 1965
+    },
+    {
+      "epoch": 1.8346530041918956,
+      "grad_norm": 0.44407746662226144,
+      "learning_rate": 3.0062143966856555e-05,
+      "loss": 0.3563,
+      "step": 1970
+    },
+    {
+      "epoch": 1.839310666045645,
+      "grad_norm": 0.4867694541510763,
+      "learning_rate": 2.9997410668047644e-05,
+      "loss": 0.362,
+      "step": 1975
+    },
+    {
+      "epoch": 1.8439683278993946,
+      "grad_norm": 0.40475826775733315,
+      "learning_rate": 2.9932677369238737e-05,
+      "loss": 0.3543,
+      "step": 1980
+    },
+    {
+      "epoch": 1.848625989753144,
+      "grad_norm": 0.439099325394701,
+      "learning_rate": 2.9867944070429834e-05,
+      "loss": 0.3683,
+      "step": 1985
+    },
+    {
+      "epoch": 1.8532836516068933,
+      "grad_norm": 0.45176402716808195,
+      "learning_rate": 2.980321077162092e-05,
+      "loss": 0.3627,
+      "step": 1990
+    },
+    {
+      "epoch": 1.8579413134606426,
+      "grad_norm": 0.39408125423927187,
+      "learning_rate": 2.9738477472812016e-05,
+      "loss": 0.3499,
+      "step": 1995
+    },
+    {
+      "epoch": 1.8625989753143921,
+      "grad_norm": 0.37626176509085807,
+      "learning_rate": 2.967374417400311e-05,
+      "loss": 0.3594,
+      "step": 2000
+    },
+    {
+      "epoch": 1.8672566371681416,
+      "grad_norm": 0.481992098530276,
+      "learning_rate": 2.96090108751942e-05,
+      "loss": 0.3655,
+      "step": 2005
+    },
+    {
+      "epoch": 1.871914299021891,
+      "grad_norm": 0.5134713155275991,
+      "learning_rate": 2.9544277576385292e-05,
+      "loss": 0.3577,
+      "step": 2010
+    },
+    {
+      "epoch": 1.8765719608756404,
+      "grad_norm": 0.45283837936405347,
+      "learning_rate": 2.9479544277576388e-05,
+      "loss": 0.3674,
+      "step": 2015
+    },
+    {
+      "epoch": 1.8812296227293899,
+      "grad_norm": 0.5331456148592331,
+      "learning_rate": 2.941481097876748e-05,
+      "loss": 0.3547,
+      "step": 2020
+    },
+    {
+      "epoch": 1.8858872845831391,
+      "grad_norm": 0.39483636058222715,
+      "learning_rate": 2.935007767995857e-05,
+      "loss": 0.3587,
+      "step": 2025
+    },
+    {
+      "epoch": 1.8905449464368886,
+      "grad_norm": 0.4169841661038559,
+      "learning_rate": 2.9285344381149664e-05,
+      "loss": 0.3562,
+      "step": 2030
+    },
+    {
+      "epoch": 1.8952026082906381,
+      "grad_norm": 0.43263533560964307,
+      "learning_rate": 2.922061108234076e-05,
+      "loss": 0.3539,
+      "step": 2035
+    },
+    {
+      "epoch": 1.8998602701443876,
+      "grad_norm": 1.0528416955125015,
+      "learning_rate": 2.915587778353185e-05,
+      "loss": 0.3656,
+      "step": 2040
+    },
+    {
+      "epoch": 1.904517931998137,
+      "grad_norm": 0.46126800798411016,
+      "learning_rate": 2.9091144484722943e-05,
+      "loss": 0.3663,
+      "step": 2045
+    },
+    {
+      "epoch": 1.9091755938518864,
+      "grad_norm": 0.439289961393162,
+      "learning_rate": 2.902641118591404e-05,
+      "loss": 0.3611,
+      "step": 2050
+    },
+    {
+      "epoch": 1.9138332557056357,
+      "grad_norm": 0.4075534808040331,
+      "learning_rate": 2.8961677887105125e-05,
+      "loss": 0.3595,
+      "step": 2055
+    },
+    {
+      "epoch": 1.9184909175593852,
+      "grad_norm": 0.44070429368587694,
+      "learning_rate": 2.889694458829622e-05,
+      "loss": 0.3554,
+      "step": 2060
+    },
+    {
+      "epoch": 1.9231485794131347,
+      "grad_norm": 0.4205873258048559,
+      "learning_rate": 2.8832211289487315e-05,
+      "loss": 0.3575,
+      "step": 2065
+    },
+    {
+      "epoch": 1.9278062412668842,
+      "grad_norm": 0.43443666788930957,
+      "learning_rate": 2.8767477990678404e-05,
+      "loss": 0.3598,
+      "step": 2070
+    },
+    {
+      "epoch": 1.9324639031206334,
+      "grad_norm": 0.4455711783118789,
+      "learning_rate": 2.8702744691869497e-05,
+      "loss": 0.3591,
+      "step": 2075
+    },
+    {
+      "epoch": 1.9371215649743827,
+      "grad_norm": 0.4633217530624874,
+      "learning_rate": 2.8638011393060594e-05,
+      "loss": 0.352,
+      "step": 2080
+    },
+    {
+      "epoch": 1.9417792268281322,
+      "grad_norm": 0.5993966669585514,
+      "learning_rate": 2.8573278094251683e-05,
+      "loss": 0.3624,
+      "step": 2085
+    },
+    {
+      "epoch": 1.9464368886818817,
+      "grad_norm": 0.47925709795843613,
+      "learning_rate": 2.8508544795442776e-05,
+      "loss": 0.347,
+      "step": 2090
+    },
+    {
+      "epoch": 1.9510945505356312,
+      "grad_norm": 0.47665074915536965,
+      "learning_rate": 2.8443811496633873e-05,
+      "loss": 0.3548,
+      "step": 2095
+    },
+    {
+      "epoch": 1.9557522123893807,
+      "grad_norm": 0.40675691505800643,
+      "learning_rate": 2.837907819782496e-05,
+      "loss": 0.3562,
+      "step": 2100
+    },
+    {
+      "epoch": 1.96040987424313,
+      "grad_norm": 0.4625409302809586,
+      "learning_rate": 2.8314344899016055e-05,
+      "loss": 0.3644,
+      "step": 2105
+    },
+    {
+      "epoch": 1.9650675360968792,
+      "grad_norm": 0.5689003662588716,
+      "learning_rate": 2.8249611600207148e-05,
+      "loss": 0.3484,
+      "step": 2110
+    },
+    {
+      "epoch": 1.9697251979506287,
+      "grad_norm": 0.5279179832753244,
+      "learning_rate": 2.8184878301398244e-05,
+      "loss": 0.3563,
+      "step": 2115
+    },
+    {
+      "epoch": 1.9743828598043782,
+      "grad_norm": 0.4038383583754883,
+      "learning_rate": 2.812014500258933e-05,
+      "loss": 0.3593,
+      "step": 2120
+    },
+    {
+      "epoch": 1.9790405216581277,
+      "grad_norm": 0.4094844987063613,
+      "learning_rate": 2.8055411703780427e-05,
+      "loss": 0.3547,
+      "step": 2125
+    },
+    {
+      "epoch": 1.983698183511877,
+      "grad_norm": 0.40606227898532676,
+      "learning_rate": 2.799067840497152e-05,
+      "loss": 0.3564,
+      "step": 2130
+    },
+    {
+      "epoch": 1.9883558453656265,
+      "grad_norm": 0.39391611988157177,
+      "learning_rate": 2.792594510616261e-05,
+      "loss": 0.3628,
+      "step": 2135
+    },
+    {
+      "epoch": 1.9930135072193758,
+      "grad_norm": 0.5333474858390276,
+      "learning_rate": 2.7861211807353703e-05,
+      "loss": 0.3614,
+      "step": 2140
+    },
+    {
+      "epoch": 1.9976711690731253,
+      "grad_norm": 0.4886820775711789,
+      "learning_rate": 2.77964785085448e-05,
+      "loss": 0.3619,
+      "step": 2145
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 4292,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.8391372660484342e+18,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-2148/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e05f31de2ff1ea05b6b40008fe030314954a34489cd1e304ab649211e4763d9
+size 7416

checkpoint-2148/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)