Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+checkpoint-241/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

added_tokens.json

@@ -0,0 +1,16 @@
+{
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}"
+}

checkpoint-241/added_tokens.json

@@ -0,0 +1,16 @@
+{
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-241/chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}"
+}

checkpoint-241/config.json

@@ -0,0 +1,48 @@
+{
+  "_name_or_path": "Qwen/Qwen2-VL-7B-Instruct",
+  "architectures": [
+    "Qwen2VLForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "image_token_id": 151655,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2_vl",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "mrope_section": [
+      16,
+      24,
+      24
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "in_chans": 3,
+    "model_type": "qwen2_vl",
+    "spatial_patch_size": 14,
+    "torch_dtype": "float32"
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "vision_token_id": 151654,
+  "vocab_size": 152064
+}

checkpoint-241/generation_config.json

@@ -0,0 +1,15 @@
+{
+  "attn_implementation": "flash_attention_2",
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.01,
+  "top_k": 1,
+  "top_p": 0.001,
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false
+}

checkpoint-241/global_step241/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:b6175182be3db7670a743495676010167ea180fb95f41098b4629bb54af79eeb
+size 3109271152

checkpoint-241/global_step241/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:f52beb8430287d1bd08d036aa74020e4b1aa7d823653a125094382c863ecd24c
+size 3109271152

checkpoint-241/global_step241/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:581a24ee5a1cc0a6361123ec5e96f4d2f597626ee4086a1c7f04cf5d4297af98
+size 3109271152

checkpoint-241/global_step241/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b333ae27b9e7075a0e1fe82a9a868b5de6c3167736b8ff5809418567d7cbd72
+size 3109271152

checkpoint-241/global_step241/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23dd9eae7b64c1da31902c95398f427cef0797f13e067b433e4f0c38c89bc3cb
+size 3109271152

checkpoint-241/global_step241/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:c5737fd8145a3ce07926ce288d370f71f16b4da5080394217687d46f93a7bd31
+size 3109271152

checkpoint-241/global_step241/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:412441d786bf76fd1db14eb8621b6498816f059d0e6eeba9652540371f4f4490
+size 3109271152

checkpoint-241/global_step241/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86d0e08704784160648ecc3c23f84bbd436cd20a23c4c34d4ff3d2a8e1deccb9
+size 3109271152

checkpoint-241/global_step241/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84d1d91fe2b33b0971ca6c13e8d5495c3db3593617f5be3cb6cda7d32a624231
+size 345056

checkpoint-241/global_step241/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35fd53e6b9cc510ee72f496b33131276b20ef67270d958eaa5e3a5210f412520
+size 345056

checkpoint-241/global_step241/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:acdb3da4b70fcce7404470e120ad0d4f44b39c3ba2e31b3ca75c62edebd5ef32
+size 345056

checkpoint-241/global_step241/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16effa550ccb979230cf8966b46c566412938839228f102552b375da75f2d703
+size 345056

checkpoint-241/global_step241/zero_pp_rank_4_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d75f9a3138f2cc46ca3e0c7e76348590882a325908ef095b2c3ea67ecfbdf1a6
+size 345056

checkpoint-241/global_step241/zero_pp_rank_5_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5dec730d4aa3558036407592b675673a4443fcc0936fc8a34a9246797b27e4c5
+size 345056

checkpoint-241/global_step241/zero_pp_rank_6_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9dd69758c3b17c6f1198f6313d770ae942f8343205511f18e3eee510e793089
+size 345056

checkpoint-241/global_step241/zero_pp_rank_7_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a81293956ad1a4d39670141ed1758953d10c3387978343a4a14fabeae52bde9
+size 345056

checkpoint-241/latest

@@ -0,0 +1 @@
+global_step241

checkpoint-241/model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18a2c8cf820035dc21186d653c096ca29364d2b91ce00a1260c5a7f53b75f4b1
+size 4966659944

checkpoint-241/model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c5862c721a0d556f5ba37db50aeed7a83228601144d832958b1c86fea9dadd3
+size 4991495816

checkpoint-241/model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ee35930f0e906c339f73759a2544adc0abb435ca8c11685da0c615c2d355e95
+size 4932751040

checkpoint-241/model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4db9563b9e5a1229fd773d02c4a06700355891f55a2038f8febd0e72d3347ea
+size 1691924384

checkpoint-241/model.safetensors.index.json

@@ -0,0 +1,737 @@
+{
+  "metadata": {
+    "total_size": 16582751232
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00003-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-00003-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00003-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.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.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00004-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-00004-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-00004-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00002-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-00002-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00002-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-00002-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-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.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",
+    "visual.blocks.0.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.0.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.1.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.10.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.11.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.12.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.13.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.14.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.15.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.16.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.17.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.18.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.19.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.2.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.20.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.21.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.22.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.23.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.24.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.25.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.26.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.27.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.28.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.29.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.3.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.30.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.31.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.4.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.5.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.6.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.7.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.8.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.norm1.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.norm1.weight": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.norm2.bias": "model-00001-of-00004.safetensors",
+    "visual.blocks.9.norm2.weight": "model-00001-of-00004.safetensors",
+    "visual.merger.ln_q.bias": "model-00001-of-00004.safetensors",
+    "visual.merger.ln_q.weight": "model-00001-of-00004.safetensors",
+    "visual.merger.mlp.0.bias": "model-00001-of-00004.safetensors",
+    "visual.merger.mlp.0.weight": "model-00001-of-00004.safetensors",
+    "visual.merger.mlp.2.bias": "model-00001-of-00004.safetensors",
+    "visual.merger.mlp.2.weight": "model-00001-of-00004.safetensors",
+    "visual.patch_embed.proj.weight": "model-00001-of-00004.safetensors"
+  }
+}

checkpoint-241/preprocessor_config.json

@@ -0,0 +1,29 @@
+{
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "Qwen2VLImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "max_pixels": 2359296,
+  "merge_size": 2,
+  "min_pixels": 3136,
+  "patch_size": 14,
+  "processor_class": "Qwen2VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "max_pixels": 12845056,
+    "min_pixels": 3136
+  },
+  "temporal_patch_size": 2
+}

checkpoint-241/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d403e954abc89e96308d8637d290658695b5e101f292b3e6dd7e3c2752b9424
+size 15984

checkpoint-241/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5eed9c892f40db2d9f0c11b64a1f2b812e420a5b54706735c68df948a4eb8a22
+size 15984

checkpoint-241/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84586131ba6d790bc95a762a9807c4782575037ebb04e4e97b5b158f035757f9
+size 15984

checkpoint-241/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:844e99f43dd145609e9beb94b89ca061ae6d08ca575716cc656424a6ed8c3fdf
+size 15984

checkpoint-241/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b683dc40255503d6ae82080494d4e0c57d3220af114ee73eae892ee98c912a70
+size 15984

checkpoint-241/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed16a9e8519510ad538724e07bf79f634447334198c717f71ade7f6da0e5b3be
+size 15984

checkpoint-241/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5abc637c74b84025333f13d2156dd89ec3b470f5f3e9f4df457f9f9f57f294ae
+size 15984

checkpoint-241/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02abfe3f8324ef502df71ba1ca3e755af045337592cbbd52990ba26e6b118de6
+size 15984

checkpoint-241/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39728186492e883fc14019879a5303128fbcf023c6d47c47abd9636d8f85c1d3
+size 1064

checkpoint-241/special_tokens_map.json

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

checkpoint-241/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88a3a6fcb80132f76da8aa40cdc3fccd7e5d8468ef15421f5b0c2715e85217d2
+size 11420538

checkpoint-241/tokenizer_config.json

@@ -0,0 +1,145 @@
+{
+  "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
+    }
+  },
+  "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": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "processor_class": "Qwen2VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-241/trainer_state.json

@@ -0,0 +1,3166 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 241,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "completion_length": 137.0703125,
+      "epoch": 0.004149377593360996,
+      "grad_norm": 39.23092968312426,
+      "kl": 0.0,
+      "learning_rate": 9.95850622406639e-07,
+      "loss": -0.0,
+      "reward": 0.3203125,
+      "reward_std": 0.3337579667568207,
+      "rewards/accuracy_reward": 0.1953125,
+      "rewards/format_reward": 0.125,
+      "step": 1
+    },
+    {
+      "completion_length": 132.765625,
+      "epoch": 0.008298755186721992,
+      "grad_norm": 22.785772433442506,
+      "kl": 0.0020294189453125,
+      "learning_rate": 9.91701244813278e-07,
+      "loss": 0.0001,
+      "reward": 0.38671875,
+      "reward_std": 0.3885350227355957,
+      "rewards/accuracy_reward": 0.20703125,
+      "rewards/format_reward": 0.1796875,
+      "step": 2
+    },
+    {
+      "completion_length": 150.3828125,
+      "epoch": 0.012448132780082987,
+      "grad_norm": 46.167032159328485,
+      "kl": 0.0030059814453125,
+      "learning_rate": 9.87551867219917e-07,
+      "loss": 0.0001,
+      "reward": 0.38671875,
+      "reward_std": 0.4264633357524872,
+      "rewards/accuracy_reward": 0.109375,
+      "rewards/format_reward": 0.27734375,
+      "step": 3
+    },
+    {
+      "completion_length": 132.20703125,
+      "epoch": 0.016597510373443983,
+      "grad_norm": 30.388246262465664,
+      "kl": 0.00701904296875,
+      "learning_rate": 9.83402489626556e-07,
+      "loss": 0.0003,
+      "reward": 0.62890625,
+      "reward_std": 0.568841814994812,
+      "rewards/accuracy_reward": 0.23046875,
+      "rewards/format_reward": 0.3984375,
+      "step": 4
+    },
+    {
+      "completion_length": 134.53515625,
+      "epoch": 0.02074688796680498,
+      "grad_norm": 9.020739859426017,
+      "kl": 0.0101318359375,
+      "learning_rate": 9.792531120331948e-07,
+      "loss": 0.0004,
+      "reward": 0.734375,
+      "reward_std": 0.49411386251449585,
+      "rewards/accuracy_reward": 0.1640625,
+      "rewards/format_reward": 0.5703125,
+      "step": 5
+    },
+    {
+      "completion_length": 144.62890625,
+      "epoch": 0.024896265560165973,
+      "grad_norm": 15.42225433374535,
+      "kl": 0.015869140625,
+      "learning_rate": 9.75103734439834e-07,
+      "loss": 0.0006,
+      "reward": 0.88671875,
+      "reward_std": 0.4697031080722809,
+      "rewards/accuracy_reward": 0.2109375,
+      "rewards/format_reward": 0.67578125,
+      "step": 6
+    },
+    {
+      "completion_length": 141.0390625,
+      "epoch": 0.029045643153526972,
+      "grad_norm": 40.46833833863793,
+      "kl": 0.01409912109375,
+      "learning_rate": 9.70954356846473e-07,
+      "loss": 0.0006,
+      "reward": 0.90625,
+      "reward_std": 0.5133033990859985,
+      "rewards/accuracy_reward": 0.23828125,
+      "rewards/format_reward": 0.66796875,
+      "step": 7
+    },
+    {
+      "completion_length": 132.3828125,
+      "epoch": 0.03319502074688797,
+      "grad_norm": 11.62408553819753,
+      "kl": 0.017578125,
+      "learning_rate": 9.66804979253112e-07,
+      "loss": 0.0007,
+      "reward": 0.9140625,
+      "reward_std": 0.4123005270957947,
+      "rewards/accuracy_reward": 0.22265625,
+      "rewards/format_reward": 0.69140625,
+      "step": 8
+    },
+    {
+      "completion_length": 127.2265625,
+      "epoch": 0.03734439834024896,
+      "grad_norm": 18.357863148161705,
+      "kl": 0.0189208984375,
+      "learning_rate": 9.62655601659751e-07,
+      "loss": 0.0008,
+      "reward": 0.90625,
+      "reward_std": 0.3882066607475281,
+      "rewards/accuracy_reward": 0.125,
+      "rewards/format_reward": 0.78125,
+      "step": 9
+    },
+    {
+      "completion_length": 123.10546875,
+      "epoch": 0.04149377593360996,
+      "grad_norm": 31.541022539135124,
+      "kl": 0.0218505859375,
+      "learning_rate": 9.5850622406639e-07,
+      "loss": 0.0009,
+      "reward": 1.08984375,
+      "reward_std": 0.339883029460907,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 0.85546875,
+      "step": 10
+    },
+    {
+      "completion_length": 106.26171875,
+      "epoch": 0.04564315352697095,
+      "grad_norm": 17.562568127530522,
+      "kl": 0.031494140625,
+      "learning_rate": 9.54356846473029e-07,
+      "loss": 0.0013,
+      "reward": 1.0546875,
+      "reward_std": 0.221227765083313,
+      "rewards/accuracy_reward": 0.109375,
+      "rewards/format_reward": 0.9453125,
+      "step": 11
+    },
+    {
+      "completion_length": 103.5546875,
+      "epoch": 0.04979253112033195,
+      "grad_norm": 9.276928026993419,
+      "kl": 0.031982421875,
+      "learning_rate": 9.50207468879668e-07,
+      "loss": 0.0013,
+      "reward": 1.09375,
+      "reward_std": 0.2309482991695404,
+      "rewards/accuracy_reward": 0.13671875,
+      "rewards/format_reward": 0.95703125,
+      "step": 12
+    },
+    {
+      "completion_length": 108.484375,
+      "epoch": 0.05394190871369295,
+      "grad_norm": 5.118367832344082,
+      "kl": 0.033447265625,
+      "learning_rate": 9.460580912863069e-07,
+      "loss": 0.0013,
+      "reward": 1.2421875,
+      "reward_std": 0.3411697447299957,
+      "rewards/accuracy_reward": 0.27734375,
+      "rewards/format_reward": 0.96484375,
+      "step": 13
+    },
+    {
+      "completion_length": 116.8515625,
+      "epoch": 0.058091286307053944,
+      "grad_norm": 12.049394415121267,
+      "kl": 0.03857421875,
+      "learning_rate": 9.41908713692946e-07,
+      "loss": 0.0015,
+      "reward": 1.1328125,
+      "reward_std": 0.21509549021720886,
+      "rewards/accuracy_reward": 0.18359375,
+      "rewards/format_reward": 0.94921875,
+      "step": 14
+    },
+    {
+      "completion_length": 107.8671875,
+      "epoch": 0.06224066390041494,
+      "grad_norm": 13.769852298124052,
+      "kl": 0.036376953125,
+      "learning_rate": 9.377593360995851e-07,
+      "loss": 0.0015,
+      "reward": 1.0390625,
+      "reward_std": 0.2277078926563263,
+      "rewards/accuracy_reward": 0.09765625,
+      "rewards/format_reward": 0.94140625,
+      "step": 15
+    },
+    {
+      "completion_length": 111.6171875,
+      "epoch": 0.06639004149377593,
+      "grad_norm": 45.043975182978954,
+      "kl": 0.359375,
+      "learning_rate": 9.33609958506224e-07,
+      "loss": 0.0143,
+      "reward": 1.0234375,
+      "reward_std": 0.263639897108078,
+      "rewards/accuracy_reward": 0.08984375,
+      "rewards/format_reward": 0.93359375,
+      "step": 16
+    },
+    {
+      "completion_length": 103.00390625,
+      "epoch": 0.07053941908713693,
+      "grad_norm": 3.010533753288046,
+      "kl": 0.04443359375,
+      "learning_rate": 9.294605809128631e-07,
+      "loss": 0.0018,
+      "reward": 1.08203125,
+      "reward_std": 0.23664388060569763,
+      "rewards/accuracy_reward": 0.109375,
+      "rewards/format_reward": 0.97265625,
+      "step": 17
+    },
+    {
+      "completion_length": 106.34375,
+      "epoch": 0.07468879668049792,
+      "grad_norm": 7.87354723018806,
+      "kl": 0.06884765625,
+      "learning_rate": 9.25311203319502e-07,
+      "loss": 0.0028,
+      "reward": 1.1953125,
+      "reward_std": 0.2938428819179535,
+      "rewards/accuracy_reward": 0.2109375,
+      "rewards/format_reward": 0.984375,
+      "step": 18
+    },
+    {
+      "completion_length": 94.46484375,
+      "epoch": 0.07883817427385892,
+      "grad_norm": 10.041019792600245,
+      "kl": 0.055419921875,
+      "learning_rate": 9.21161825726141e-07,
+      "loss": 0.0022,
+      "reward": 1.21875,
+      "reward_std": 0.2210792601108551,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.9765625,
+      "step": 19
+    },
+    {
+      "completion_length": 92.640625,
+      "epoch": 0.08298755186721991,
+      "grad_norm": 7.905738191363626,
+      "kl": 0.05615234375,
+      "learning_rate": 9.170124481327801e-07,
+      "loss": 0.0022,
+      "reward": 1.11328125,
+      "reward_std": 0.21515560150146484,
+      "rewards/accuracy_reward": 0.140625,
+      "rewards/format_reward": 0.97265625,
+      "step": 20
+    },
+    {
+      "completion_length": 90.3984375,
+      "epoch": 0.08713692946058091,
+      "grad_norm": 5.157524094818503,
+      "kl": 0.06298828125,
+      "learning_rate": 9.12863070539419e-07,
+      "loss": 0.0025,
+      "reward": 1.26171875,
+      "reward_std": 0.2477707862854004,
+      "rewards/accuracy_reward": 0.26953125,
+      "rewards/format_reward": 0.9921875,
+      "step": 21
+    },
+    {
+      "completion_length": 97.3203125,
+      "epoch": 0.0912863070539419,
+      "grad_norm": 7.010674413524611,
+      "kl": 0.072265625,
+      "learning_rate": 9.087136929460581e-07,
+      "loss": 0.0029,
+      "reward": 1.23046875,
+      "reward_std": 0.20519250631332397,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.98828125,
+      "step": 22
+    },
+    {
+      "completion_length": 93.70703125,
+      "epoch": 0.0954356846473029,
+      "grad_norm": 895.4823242055704,
+      "kl": 4.6875,
+      "learning_rate": 9.04564315352697e-07,
+      "loss": 0.1871,
+      "reward": 1.1953125,
+      "reward_std": 0.2628646492958069,
+      "rewards/accuracy_reward": 0.2265625,
+      "rewards/format_reward": 0.96875,
+      "step": 23
+    },
+    {
+      "completion_length": 92.921875,
+      "epoch": 0.0995850622406639,
+      "grad_norm": 9.141621520119115,
+      "kl": 0.087890625,
+      "learning_rate": 9.004149377593361e-07,
+      "loss": 0.0035,
+      "reward": 1.140625,
+      "reward_std": 0.23703515529632568,
+      "rewards/accuracy_reward": 0.1640625,
+      "rewards/format_reward": 0.9765625,
+      "step": 24
+    },
+    {
+      "completion_length": 89.5859375,
+      "epoch": 0.1037344398340249,
+      "grad_norm": 11.592503375213493,
+      "kl": 0.0888671875,
+      "learning_rate": 8.96265560165975e-07,
+      "loss": 0.0036,
+      "reward": 1.2578125,
+      "reward_std": 0.2782890200614929,
+      "rewards/accuracy_reward": 0.26953125,
+      "rewards/format_reward": 0.98828125,
+      "step": 25
+    },
+    {
+      "completion_length": 92.765625,
+      "epoch": 0.1078838174273859,
+      "grad_norm": 4.844037513428872,
+      "kl": 0.08740234375,
+      "learning_rate": 8.921161825726141e-07,
+      "loss": 0.0035,
+      "reward": 1.21484375,
+      "reward_std": 0.3296912610530853,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 0.98046875,
+      "step": 26
+    },
+    {
+      "completion_length": 77.96484375,
+      "epoch": 0.11203319502074689,
+      "grad_norm": 28.980687466181028,
+      "kl": 0.11279296875,
+      "learning_rate": 8.879668049792531e-07,
+      "loss": 0.0045,
+      "reward": 1.2421875,
+      "reward_std": 0.24264347553253174,
+      "rewards/accuracy_reward": 0.2578125,
+      "rewards/format_reward": 0.984375,
+      "step": 27
+    },
+    {
+      "completion_length": 72.87109375,
+      "epoch": 0.11618257261410789,
+      "grad_norm": 7972.551755677257,
+      "kl": 2.359375,
+      "learning_rate": 8.83817427385892e-07,
+      "loss": 0.0942,
+      "reward": 1.25390625,
+      "reward_std": 0.26447463035583496,
+      "rewards/accuracy_reward": 0.26953125,
+      "rewards/format_reward": 0.984375,
+      "step": 28
+    },
+    {
+      "completion_length": 80.0234375,
+      "epoch": 0.12033195020746888,
+      "grad_norm": 148541.21316902922,
+      "kl": 102.0,
+      "learning_rate": 8.796680497925311e-07,
+      "loss": 4.0946,
+      "reward": 1.21484375,
+      "reward_std": 0.2021351009607315,
+      "rewards/accuracy_reward": 0.22265625,
+      "rewards/format_reward": 0.9921875,
+      "step": 29
+    },
+    {
+      "completion_length": 71.890625,
+      "epoch": 0.12448132780082988,
+      "grad_norm": 155.11184072334507,
+      "kl": 0.154296875,
+      "learning_rate": 8.7551867219917e-07,
+      "loss": 0.0062,
+      "reward": 1.2421875,
+      "reward_std": 0.33442190289497375,
+      "rewards/accuracy_reward": 0.28125,
+      "rewards/format_reward": 0.9609375,
+      "step": 30
+    },
+    {
+      "completion_length": 72.9765625,
+      "epoch": 0.12863070539419086,
+      "grad_norm": 6.669125194641157,
+      "kl": 0.1435546875,
+      "learning_rate": 8.713692946058091e-07,
+      "loss": 0.0058,
+      "reward": 1.25390625,
+      "reward_std": 0.26459574699401855,
+      "rewards/accuracy_reward": 0.25390625,
+      "rewards/format_reward": 1.0,
+      "step": 31
+    },
+    {
+      "completion_length": 92.39453125,
+      "epoch": 0.13278008298755187,
+      "grad_norm": 8.594154520894591,
+      "kl": 0.12060546875,
+      "learning_rate": 8.672199170124481e-07,
+      "loss": 0.0048,
+      "reward": 1.296875,
+      "reward_std": 0.26187676191329956,
+      "rewards/accuracy_reward": 0.31640625,
+      "rewards/format_reward": 0.98046875,
+      "step": 32
+    },
+    {
+      "completion_length": 78.91796875,
+      "epoch": 0.13692946058091288,
+      "grad_norm": 7.555816953155662,
+      "kl": 0.158203125,
+      "learning_rate": 8.630705394190871e-07,
+      "loss": 0.0063,
+      "reward": 1.203125,
+      "reward_std": 0.25460314750671387,
+      "rewards/accuracy_reward": 0.21484375,
+      "rewards/format_reward": 0.98828125,
+      "step": 33
+    },
+    {
+      "completion_length": 73.20703125,
+      "epoch": 0.14107883817427386,
+      "grad_norm": 19.83628645685219,
+      "kl": 0.1591796875,
+      "learning_rate": 8.589211618257261e-07,
+      "loss": 0.0064,
+      "reward": 1.28515625,
+      "reward_std": 0.3227723240852356,
+      "rewards/accuracy_reward": 0.30078125,
+      "rewards/format_reward": 0.984375,
+      "step": 34
+    },
+    {
+      "completion_length": 73.3515625,
+      "epoch": 0.14522821576763487,
+      "grad_norm": 8.311376906204279,
+      "kl": 0.2197265625,
+      "learning_rate": 8.54771784232365e-07,
+      "loss": 0.0088,
+      "reward": 1.1953125,
+      "reward_std": 0.2911452054977417,
+      "rewards/accuracy_reward": 0.21875,
+      "rewards/format_reward": 0.9765625,
+      "step": 35
+    },
+    {
+      "completion_length": 81.99609375,
+      "epoch": 0.14937759336099585,
+      "grad_norm": 10.43036199186392,
+      "kl": 0.12890625,
+      "learning_rate": 8.506224066390041e-07,
+      "loss": 0.0052,
+      "reward": 1.2578125,
+      "reward_std": 0.3043610155582428,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.984375,
+      "step": 36
+    },
+    {
+      "completion_length": 73.796875,
+      "epoch": 0.15352697095435686,
+      "grad_norm": 57.24643236121668,
+      "kl": 0.1357421875,
+      "learning_rate": 8.464730290456431e-07,
+      "loss": 0.0054,
+      "reward": 1.21875,
+      "reward_std": 0.28197723627090454,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.96875,
+      "step": 37
+    },
+    {
+      "completion_length": 89.75390625,
+      "epoch": 0.15767634854771784,
+      "grad_norm": 19.872397533152718,
+      "kl": 0.1064453125,
+      "learning_rate": 8.423236514522821e-07,
+      "loss": 0.0043,
+      "reward": 1.234375,
+      "reward_std": 0.250660240650177,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 1.0,
+      "step": 38
+    },
+    {
+      "completion_length": 92.703125,
+      "epoch": 0.16182572614107885,
+      "grad_norm": 11.503069481139981,
+      "kl": 0.09326171875,
+      "learning_rate": 8.381742738589212e-07,
+      "loss": 0.0037,
+      "reward": 1.234375,
+      "reward_std": 0.34505003690719604,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.9609375,
+      "step": 39
+    },
+    {
+      "completion_length": 89.04296875,
+      "epoch": 0.16597510373443983,
+      "grad_norm": 18.044182586396225,
+      "kl": 0.103515625,
+      "learning_rate": 8.340248962655602e-07,
+      "loss": 0.0041,
+      "reward": 1.140625,
+      "reward_std": 0.28612154722213745,
+      "rewards/accuracy_reward": 0.171875,
+      "rewards/format_reward": 0.96875,
+      "step": 40
+    },
+    {
+      "completion_length": 92.12890625,
+      "epoch": 0.17012448132780084,
+      "grad_norm": 955.7098863804349,
+      "kl": 6.21875,
+      "learning_rate": 8.298755186721992e-07,
+      "loss": 0.2497,
+      "reward": 1.15625,
+      "reward_std": 0.3010457158088684,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.91015625,
+      "step": 41
+    },
+    {
+      "completion_length": 92.9375,
+      "epoch": 0.17427385892116182,
+      "grad_norm": 19789.27042408252,
+      "kl": 38.75,
+      "learning_rate": 8.257261410788381e-07,
+      "loss": 1.5507,
+      "reward": 1.17578125,
+      "reward_std": 0.24686214327812195,
+      "rewards/accuracy_reward": 0.20703125,
+      "rewards/format_reward": 0.96875,
+      "step": 42
+    },
+    {
+      "completion_length": 103.703125,
+      "epoch": 0.17842323651452283,
+      "grad_norm": 5.178711424894338,
+      "kl": 0.07666015625,
+      "learning_rate": 8.215767634854771e-07,
+      "loss": 0.0031,
+      "reward": 1.16015625,
+      "reward_std": 0.2846824824810028,
+      "rewards/accuracy_reward": 0.1875,
+      "rewards/format_reward": 0.97265625,
+      "step": 43
+    },
+    {
+      "completion_length": 96.78125,
+      "epoch": 0.1825726141078838,
+      "grad_norm": 79.81973712987309,
+      "kl": 0.2138671875,
+      "learning_rate": 8.174273858921161e-07,
+      "loss": 0.0086,
+      "reward": 1.1171875,
+      "reward_std": 0.2437995821237564,
+      "rewards/accuracy_reward": 0.1875,
+      "rewards/format_reward": 0.9296875,
+      "step": 44
+    },
+    {
+      "completion_length": 105.375,
+      "epoch": 0.18672199170124482,
+      "grad_norm": 44.03437371953703,
+      "kl": 0.1083984375,
+      "learning_rate": 8.132780082987552e-07,
+      "loss": 0.0043,
+      "reward": 1.1328125,
+      "reward_std": 0.28126898407936096,
+      "rewards/accuracy_reward": 0.171875,
+      "rewards/format_reward": 0.9609375,
+      "step": 45
+    },
+    {
+      "completion_length": 95.625,
+      "epoch": 0.1908713692946058,
+      "grad_norm": 255.29741353447812,
+      "kl": 0.2158203125,
+      "learning_rate": 8.091286307053942e-07,
+      "loss": 0.0087,
+      "reward": 1.26953125,
+      "reward_std": 0.30966654419898987,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.984375,
+      "step": 46
+    },
+    {
+      "completion_length": 90.71484375,
+      "epoch": 0.1950207468879668,
+      "grad_norm": 141.664562557487,
+      "kl": 0.28125,
+      "learning_rate": 8.049792531120332e-07,
+      "loss": 0.0112,
+      "reward": 1.10546875,
+      "reward_std": 0.2388685643672943,
+      "rewards/accuracy_reward": 0.1484375,
+      "rewards/format_reward": 0.95703125,
+      "step": 47
+    },
+    {
+      "completion_length": 104.390625,
+      "epoch": 0.1991701244813278,
+      "grad_norm": 26.547793382795557,
+      "kl": 0.0625,
+      "learning_rate": 8.008298755186722e-07,
+      "loss": 0.0025,
+      "reward": 1.17578125,
+      "reward_std": 0.3045119643211365,
+      "rewards/accuracy_reward": 0.21875,
+      "rewards/format_reward": 0.95703125,
+      "step": 48
+    },
+    {
+      "completion_length": 103.8359375,
+      "epoch": 0.2033195020746888,
+      "grad_norm": 11.71623541450062,
+      "kl": 0.06103515625,
+      "learning_rate": 7.966804979253111e-07,
+      "loss": 0.0024,
+      "reward": 1.26953125,
+      "reward_std": 0.31901323795318604,
+      "rewards/accuracy_reward": 0.3046875,
+      "rewards/format_reward": 0.96484375,
+      "step": 49
+    },
+    {
+      "completion_length": 98.74609375,
+      "epoch": 0.2074688796680498,
+      "grad_norm": 28.06161637024887,
+      "kl": 0.078125,
+      "learning_rate": 7.925311203319502e-07,
+      "loss": 0.0031,
+      "reward": 1.15234375,
+      "reward_std": 0.23013246059417725,
+      "rewards/accuracy_reward": 0.19140625,
+      "rewards/format_reward": 0.9609375,
+      "step": 50
+    },
+    {
+      "completion_length": 106.44140625,
+      "epoch": 0.21161825726141079,
+      "grad_norm": 8.385727882590416,
+      "kl": 0.057373046875,
+      "learning_rate": 7.883817427385891e-07,
+      "loss": 0.0023,
+      "reward": 1.1953125,
+      "reward_std": 0.2811848223209381,
+      "rewards/accuracy_reward": 0.21484375,
+      "rewards/format_reward": 0.98046875,
+      "step": 51
+    },
+    {
+      "completion_length": 102.765625,
+      "epoch": 0.2157676348547718,
+      "grad_norm": 6.2531692527733425,
+      "kl": 0.06787109375,
+      "learning_rate": 7.842323651452282e-07,
+      "loss": 0.0027,
+      "reward": 1.34765625,
+      "reward_std": 0.2933124899864197,
+      "rewards/accuracy_reward": 0.375,
+      "rewards/format_reward": 0.97265625,
+      "step": 52
+    },
+    {
+      "completion_length": 108.328125,
+      "epoch": 0.21991701244813278,
+      "grad_norm": 3.8970282813938035,
+      "kl": 0.05859375,
+      "learning_rate": 7.800829875518672e-07,
+      "loss": 0.0023,
+      "reward": 1.19921875,
+      "reward_std": 0.27631211280822754,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.95703125,
+      "step": 53
+    },
+    {
+      "completion_length": 115.71875,
+      "epoch": 0.22406639004149378,
+      "grad_norm": 4.0861620229361435,
+      "kl": 0.06396484375,
+      "learning_rate": 7.759336099585062e-07,
+      "loss": 0.0026,
+      "reward": 1.21875,
+      "reward_std": 0.32264500856399536,
+      "rewards/accuracy_reward": 0.26953125,
+      "rewards/format_reward": 0.94921875,
+      "step": 54
+    },
+    {
+      "completion_length": 109.42578125,
+      "epoch": 0.22821576763485477,
+      "grad_norm": 2.2229514837482967,
+      "kl": 0.0576171875,
+      "learning_rate": 7.717842323651453e-07,
+      "loss": 0.0023,
+      "reward": 1.1875,
+      "reward_std": 0.2718299627304077,
+      "rewards/accuracy_reward": 0.22265625,
+      "rewards/format_reward": 0.96484375,
+      "step": 55
+    },
+    {
+      "completion_length": 114.6171875,
+      "epoch": 0.23236514522821577,
+      "grad_norm": 6.044557856528395,
+      "kl": 0.051513671875,
+      "learning_rate": 7.676348547717842e-07,
+      "loss": 0.0021,
+      "reward": 1.28125,
+      "reward_std": 0.32618969678878784,
+      "rewards/accuracy_reward": 0.3125,
+      "rewards/format_reward": 0.96875,
+      "step": 56
+    },
+    {
+      "completion_length": 104.9921875,
+      "epoch": 0.23651452282157676,
+      "grad_norm": 3.7345492151935566,
+      "kl": 0.06298828125,
+      "learning_rate": 7.634854771784232e-07,
+      "loss": 0.0025,
+      "reward": 1.33203125,
+      "reward_std": 0.263650119304657,
+      "rewards/accuracy_reward": 0.359375,
+      "rewards/format_reward": 0.97265625,
+      "step": 57
+    },
+    {
+      "completion_length": 117.24609375,
+      "epoch": 0.24066390041493776,
+      "grad_norm": 9.681979509468285,
+      "kl": 0.054443359375,
+      "learning_rate": 7.593360995850621e-07,
+      "loss": 0.0022,
+      "reward": 1.125,
+      "reward_std": 0.26918965578079224,
+      "rewards/accuracy_reward": 0.18359375,
+      "rewards/format_reward": 0.94140625,
+      "step": 58
+    },
+    {
+      "completion_length": 110.26953125,
+      "epoch": 0.24481327800829875,
+      "grad_norm": 11.501649552098742,
+      "kl": 0.08251953125,
+      "learning_rate": 7.551867219917012e-07,
+      "loss": 0.0033,
+      "reward": 1.2109375,
+      "reward_std": 0.22527292370796204,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.9609375,
+      "step": 59
+    },
+    {
+      "completion_length": 108.19921875,
+      "epoch": 0.24896265560165975,
+      "grad_norm": 9.965135528231466,
+      "kl": 0.06591796875,
+      "learning_rate": 7.510373443983403e-07,
+      "loss": 0.0026,
+      "reward": 1.2734375,
+      "reward_std": 0.3159399628639221,
+      "rewards/accuracy_reward": 0.328125,
+      "rewards/format_reward": 0.9453125,
+      "step": 60
+    },
+    {
+      "completion_length": 121.76171875,
+      "epoch": 0.25311203319502074,
+      "grad_norm": 1.8431613725044658,
+      "kl": 0.054931640625,
+      "learning_rate": 7.468879668049792e-07,
+      "loss": 0.0022,
+      "reward": 1.23046875,
+      "reward_std": 0.30954664945602417,
+      "rewards/accuracy_reward": 0.27734375,
+      "rewards/format_reward": 0.953125,
+      "step": 61
+    },
+    {
+      "completion_length": 115.984375,
+      "epoch": 0.2572614107883817,
+      "grad_norm": 14.722550985552658,
+      "kl": 0.07763671875,
+      "learning_rate": 7.427385892116183e-07,
+      "loss": 0.0031,
+      "reward": 1.1171875,
+      "reward_std": 0.2740333080291748,
+      "rewards/accuracy_reward": 0.1875,
+      "rewards/format_reward": 0.9296875,
+      "step": 62
+    },
+    {
+      "completion_length": 122.40625,
+      "epoch": 0.26141078838174275,
+      "grad_norm": 2.3435103733008082,
+      "kl": 0.0576171875,
+      "learning_rate": 7.385892116182572e-07,
+      "loss": 0.0023,
+      "reward": 1.16015625,
+      "reward_std": 0.32881489396095276,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.9140625,
+      "step": 63
+    },
+    {
+      "completion_length": 111.37890625,
+      "epoch": 0.26556016597510373,
+      "grad_norm": 5.3464512193055205,
+      "kl": 0.0654296875,
+      "learning_rate": 7.344398340248963e-07,
+      "loss": 0.0026,
+      "reward": 1.26953125,
+      "reward_std": 0.27697107195854187,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.984375,
+      "step": 64
+    },
+    {
+      "completion_length": 109.5078125,
+      "epoch": 0.2697095435684647,
+      "grad_norm": 22.06078984271988,
+      "kl": 0.06298828125,
+      "learning_rate": 7.302904564315351e-07,
+      "loss": 0.0025,
+      "reward": 1.2109375,
+      "reward_std": 0.2712310552597046,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.96484375,
+      "step": 65
+    },
+    {
+      "completion_length": 120.81640625,
+      "epoch": 0.27385892116182575,
+      "grad_norm": 7.466388660174733,
+      "kl": 0.06298828125,
+      "learning_rate": 7.261410788381742e-07,
+      "loss": 0.0025,
+      "reward": 1.23828125,
+      "reward_std": 0.27975520491600037,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.953125,
+      "step": 66
+    },
+    {
+      "completion_length": 110.5546875,
+      "epoch": 0.27800829875518673,
+      "grad_norm": 6.507257995412741,
+      "kl": 0.06787109375,
+      "learning_rate": 7.219917012448133e-07,
+      "loss": 0.0027,
+      "reward": 1.24609375,
+      "reward_std": 0.28682631254196167,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.97265625,
+      "step": 67
+    },
+    {
+      "completion_length": 107.703125,
+      "epoch": 0.2821576763485477,
+      "grad_norm": 10.657799163428932,
+      "kl": 0.061767578125,
+      "learning_rate": 7.178423236514522e-07,
+      "loss": 0.0025,
+      "reward": 1.23046875,
+      "reward_std": 0.262653648853302,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.95703125,
+      "step": 68
+    },
+    {
+      "completion_length": 98.36328125,
+      "epoch": 0.2863070539419087,
+      "grad_norm": 7.291922232973526,
+      "kl": 0.0634765625,
+      "learning_rate": 7.136929460580913e-07,
+      "loss": 0.0025,
+      "reward": 1.265625,
+      "reward_std": 0.2502795457839966,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.9921875,
+      "step": 69
+    },
+    {
+      "completion_length": 108.15625,
+      "epoch": 0.29045643153526973,
+      "grad_norm": 2531.5651653482937,
+      "kl": 0.77734375,
+      "learning_rate": 7.095435684647303e-07,
+      "loss": 0.0311,
+      "reward": 1.1953125,
+      "reward_std": 0.3056316375732422,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 0.9609375,
+      "step": 70
+    },
+    {
+      "completion_length": 122.08203125,
+      "epoch": 0.2946058091286307,
+      "grad_norm": 7.2767653534750245,
+      "kl": 0.05517578125,
+      "learning_rate": 7.053941908713693e-07,
+      "loss": 0.0022,
+      "reward": 1.21484375,
+      "reward_std": 0.33597099781036377,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.94921875,
+      "step": 71
+    },
+    {
+      "completion_length": 112.4140625,
+      "epoch": 0.2987551867219917,
+      "grad_norm": 5.558748098185352,
+      "kl": 0.0576171875,
+      "learning_rate": 7.012448132780083e-07,
+      "loss": 0.0023,
+      "reward": 1.23046875,
+      "reward_std": 0.27381569147109985,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.98046875,
+      "step": 72
+    },
+    {
+      "completion_length": 115.1171875,
+      "epoch": 0.3029045643153527,
+      "grad_norm": 146.48646129649492,
+      "kl": 0.1416015625,
+      "learning_rate": 6.970954356846472e-07,
+      "loss": 0.0056,
+      "reward": 1.24609375,
+      "reward_std": 0.3475721478462219,
+      "rewards/accuracy_reward": 0.296875,
+      "rewards/format_reward": 0.94921875,
+      "step": 73
+    },
+    {
+      "completion_length": 105.43359375,
+      "epoch": 0.3070539419087137,
+      "grad_norm": 20.98944404533785,
+      "kl": 0.1474609375,
+      "learning_rate": 6.929460580912863e-07,
+      "loss": 0.0059,
+      "reward": 1.18359375,
+      "reward_std": 0.1996288001537323,
+      "rewards/accuracy_reward": 0.22265625,
+      "rewards/format_reward": 0.9609375,
+      "step": 74
+    },
+    {
+      "completion_length": 107.79296875,
+      "epoch": 0.3112033195020747,
+      "grad_norm": 2.8266704753608516,
+      "kl": 0.06201171875,
+      "learning_rate": 6.887966804979252e-07,
+      "loss": 0.0025,
+      "reward": 1.25,
+      "reward_std": 0.2760782241821289,
+      "rewards/accuracy_reward": 0.26953125,
+      "rewards/format_reward": 0.98046875,
+      "step": 75
+    },
+    {
+      "completion_length": 121.66796875,
+      "epoch": 0.3153526970954357,
+      "grad_norm": 77.92500399218551,
+      "kl": 0.06689453125,
+      "learning_rate": 6.846473029045643e-07,
+      "loss": 0.0027,
+      "reward": 1.24609375,
+      "reward_std": 0.30910396575927734,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.9609375,
+      "step": 76
+    },
+    {
+      "completion_length": 111.671875,
+      "epoch": 0.31950207468879666,
+      "grad_norm": 4.702893593213397,
+      "kl": 0.05224609375,
+      "learning_rate": 6.804979253112033e-07,
+      "loss": 0.0021,
+      "reward": 1.1953125,
+      "reward_std": 0.224736750125885,
+      "rewards/accuracy_reward": 0.22265625,
+      "rewards/format_reward": 0.97265625,
+      "step": 77
+    },
+    {
+      "completion_length": 100.1640625,
+      "epoch": 0.3236514522821577,
+      "grad_norm": 14.209228868230555,
+      "kl": 0.08740234375,
+      "learning_rate": 6.763485477178423e-07,
+      "loss": 0.0035,
+      "reward": 1.171875,
+      "reward_std": 0.21961188316345215,
+      "rewards/accuracy_reward": 0.19921875,
+      "rewards/format_reward": 0.97265625,
+      "step": 78
+    },
+    {
+      "completion_length": 109.49609375,
+      "epoch": 0.3278008298755187,
+      "grad_norm": 579.9463703947171,
+      "kl": 0.18359375,
+      "learning_rate": 6.721991701244814e-07,
+      "loss": 0.0074,
+      "reward": 1.19140625,
+      "reward_std": 0.3556343615055084,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.94140625,
+      "step": 79
+    },
+    {
+      "completion_length": 105.04296875,
+      "epoch": 0.33195020746887965,
+      "grad_norm": 5927.641237621651,
+      "kl": 6.4375,
+      "learning_rate": 6.680497925311202e-07,
+      "loss": 0.2576,
+      "reward": 1.25390625,
+      "reward_std": 0.337135374546051,
+      "rewards/accuracy_reward": 0.2890625,
+      "rewards/format_reward": 0.96484375,
+      "step": 80
+    },
+    {
+      "completion_length": 104.91796875,
+      "epoch": 0.3360995850622407,
+      "grad_norm": 4.577287332890265,
+      "kl": 0.068359375,
+      "learning_rate": 6.639004149377593e-07,
+      "loss": 0.0027,
+      "reward": 1.2578125,
+      "reward_std": 0.22225633263587952,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.9921875,
+      "step": 81
+    },
+    {
+      "completion_length": 110.57421875,
+      "epoch": 0.34024896265560167,
+      "grad_norm": 80.31923623358058,
+      "kl": 0.142578125,
+      "learning_rate": 6.597510373443983e-07,
+      "loss": 0.0057,
+      "reward": 1.24609375,
+      "reward_std": 0.30580008029937744,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.9609375,
+      "step": 82
+    },
+    {
+      "completion_length": 110.4140625,
+      "epoch": 0.34439834024896265,
+      "grad_norm": 3.697288528310834,
+      "kl": 0.0712890625,
+      "learning_rate": 6.556016597510373e-07,
+      "loss": 0.0029,
+      "reward": 1.25390625,
+      "reward_std": 0.2552226185798645,
+      "rewards/accuracy_reward": 0.28125,
+      "rewards/format_reward": 0.97265625,
+      "step": 83
+    },
+    {
+      "completion_length": 104.67578125,
+      "epoch": 0.34854771784232363,
+      "grad_norm": 5.229318975289039,
+      "kl": 0.06982421875,
+      "learning_rate": 6.514522821576763e-07,
+      "loss": 0.0028,
+      "reward": 1.26953125,
+      "reward_std": 0.24407757818698883,
+      "rewards/accuracy_reward": 0.29296875,
+      "rewards/format_reward": 0.9765625,
+      "step": 84
+    },
+    {
+      "completion_length": 103.8828125,
+      "epoch": 0.35269709543568467,
+      "grad_norm": 2252.8465055849706,
+      "kl": 18.75,
+      "learning_rate": 6.473029045643154e-07,
+      "loss": 0.7462,
+      "reward": 1.21875,
+      "reward_std": 0.32842758297920227,
+      "rewards/accuracy_reward": 0.23828125,
+      "rewards/format_reward": 0.98046875,
+      "step": 85
+    },
+    {
+      "completion_length": 110.5625,
+      "epoch": 0.35684647302904565,
+      "grad_norm": 4.971947739502407,
+      "kl": 0.0673828125,
+      "learning_rate": 6.431535269709544e-07,
+      "loss": 0.0027,
+      "reward": 1.2421875,
+      "reward_std": 0.28188830614089966,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.96875,
+      "step": 86
+    },
+    {
+      "completion_length": 118.6484375,
+      "epoch": 0.36099585062240663,
+      "grad_norm": 8.208637190208675,
+      "kl": 0.0634765625,
+      "learning_rate": 6.390041493775933e-07,
+      "loss": 0.0025,
+      "reward": 1.06640625,
+      "reward_std": 0.25959229469299316,
+      "rewards/accuracy_reward": 0.11328125,
+      "rewards/format_reward": 0.953125,
+      "step": 87
+    },
+    {
+      "completion_length": 96.6875,
+      "epoch": 0.3651452282157676,
+      "grad_norm": 38.459676626860556,
+      "kl": 0.1552734375,
+      "learning_rate": 6.348547717842323e-07,
+      "loss": 0.0062,
+      "reward": 1.29296875,
+      "reward_std": 0.29796260595321655,
+      "rewards/accuracy_reward": 0.34765625,
+      "rewards/format_reward": 0.9453125,
+      "step": 88
+    },
+    {
+      "completion_length": 92.484375,
+      "epoch": 0.36929460580912865,
+      "grad_norm": 11.251071020115205,
+      "kl": 0.087890625,
+      "learning_rate": 6.307053941908713e-07,
+      "loss": 0.0035,
+      "reward": 1.1328125,
+      "reward_std": 0.26419880986213684,
+      "rewards/accuracy_reward": 0.171875,
+      "rewards/format_reward": 0.9609375,
+      "step": 89
+    },
+    {
+      "completion_length": 94.71875,
+      "epoch": 0.37344398340248963,
+      "grad_norm": 73.08860324819888,
+      "kl": 0.1318359375,
+      "learning_rate": 6.265560165975103e-07,
+      "loss": 0.0053,
+      "reward": 1.24609375,
+      "reward_std": 0.28619101643562317,
+      "rewards/accuracy_reward": 0.26171875,
+      "rewards/format_reward": 0.984375,
+      "step": 90
+    },
+    {
+      "completion_length": 105.15625,
+      "epoch": 0.3775933609958506,
+      "grad_norm": 32.762748338362975,
+      "kl": 0.07763671875,
+      "learning_rate": 6.224066390041493e-07,
+      "loss": 0.0031,
+      "reward": 1.2109375,
+      "reward_std": 0.33037883043289185,
+      "rewards/accuracy_reward": 0.23828125,
+      "rewards/format_reward": 0.97265625,
+      "step": 91
+    },
+    {
+      "completion_length": 96.0703125,
+      "epoch": 0.3817427385892116,
+      "grad_norm": 5.087858102518854,
+      "kl": 0.07861328125,
+      "learning_rate": 6.182572614107884e-07,
+      "loss": 0.0031,
+      "reward": 1.2265625,
+      "reward_std": 0.3091992139816284,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.984375,
+      "step": 92
+    },
+    {
+      "completion_length": 100.95703125,
+      "epoch": 0.38589211618257263,
+      "grad_norm": 12.724022811120527,
+      "kl": 0.0830078125,
+      "learning_rate": 6.141078838174274e-07,
+      "loss": 0.0033,
+      "reward": 1.21875,
+      "reward_std": 0.29461750388145447,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.97265625,
+      "step": 93
+    },
+    {
+      "completion_length": 104.11328125,
+      "epoch": 0.3900414937759336,
+      "grad_norm": 17.194627861953816,
+      "kl": 0.08642578125,
+      "learning_rate": 6.099585062240664e-07,
+      "loss": 0.0035,
+      "reward": 1.2265625,
+      "reward_std": 0.2846662104129791,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.984375,
+      "step": 94
+    },
+    {
+      "completion_length": 92.14453125,
+      "epoch": 0.3941908713692946,
+      "grad_norm": 7.192313306035128,
+      "kl": 0.08642578125,
+      "learning_rate": 6.058091286307053e-07,
+      "loss": 0.0035,
+      "reward": 1.2734375,
+      "reward_std": 0.23374123871326447,
+      "rewards/accuracy_reward": 0.2890625,
+      "rewards/format_reward": 0.984375,
+      "step": 95
+    },
+    {
+      "completion_length": 84.6328125,
+      "epoch": 0.3983402489626556,
+      "grad_norm": 47.641365377607805,
+      "kl": 0.1025390625,
+      "learning_rate": 6.016597510373443e-07,
+      "loss": 0.0041,
+      "reward": 1.2890625,
+      "reward_std": 0.28897759318351746,
+      "rewards/accuracy_reward": 0.30859375,
+      "rewards/format_reward": 0.98046875,
+      "step": 96
+    },
+    {
+      "completion_length": 78.61328125,
+      "epoch": 0.4024896265560166,
+      "grad_norm": 253.81419290673912,
+      "kl": 1.9609375,
+      "learning_rate": 5.975103734439834e-07,
+      "loss": 0.0787,
+      "reward": 1.2734375,
+      "reward_std": 0.2571094334125519,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.98828125,
+      "step": 97
+    },
+    {
+      "completion_length": 88.37109375,
+      "epoch": 0.4066390041493776,
+      "grad_norm": 12.19756385611491,
+      "kl": 0.0986328125,
+      "learning_rate": 5.933609958506224e-07,
+      "loss": 0.0039,
+      "reward": 1.234375,
+      "reward_std": 0.2579057812690735,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.984375,
+      "step": 98
+    },
+    {
+      "completion_length": 94.3125,
+      "epoch": 0.4107883817427386,
+      "grad_norm": 8.78041276853543,
+      "kl": 0.0927734375,
+      "learning_rate": 5.892116182572614e-07,
+      "loss": 0.0037,
+      "reward": 1.18359375,
+      "reward_std": 0.32227471470832825,
+      "rewards/accuracy_reward": 0.21875,
+      "rewards/format_reward": 0.96484375,
+      "step": 99
+    },
+    {
+      "completion_length": 92.3828125,
+      "epoch": 0.4149377593360996,
+      "grad_norm": 8.518814664938573,
+      "kl": 0.0859375,
+      "learning_rate": 5.850622406639005e-07,
+      "loss": 0.0034,
+      "reward": 1.21484375,
+      "reward_std": 0.24748478829860687,
+      "rewards/accuracy_reward": 0.23046875,
+      "rewards/format_reward": 0.984375,
+      "step": 100
+    },
+    {
+      "completion_length": 85.60546875,
+      "epoch": 0.4190871369294606,
+      "grad_norm": 4.704027215633879,
+      "kl": 0.10205078125,
+      "learning_rate": 5.809128630705394e-07,
+      "loss": 0.0041,
+      "reward": 1.1796875,
+      "reward_std": 0.22647252678871155,
+      "rewards/accuracy_reward": 0.19140625,
+      "rewards/format_reward": 0.98828125,
+      "step": 101
+    },
+    {
+      "completion_length": 99.41796875,
+      "epoch": 0.42323651452282157,
+      "grad_norm": 4.680746289931085,
+      "kl": 0.08935546875,
+      "learning_rate": 5.767634854771784e-07,
+      "loss": 0.0036,
+      "reward": 1.36328125,
+      "reward_std": 0.3082295060157776,
+      "rewards/accuracy_reward": 0.37890625,
+      "rewards/format_reward": 0.984375,
+      "step": 102
+    },
+    {
+      "completion_length": 93.4453125,
+      "epoch": 0.42738589211618255,
+      "grad_norm": 6.65500574546716,
+      "kl": 0.1103515625,
+      "learning_rate": 5.726141078838173e-07,
+      "loss": 0.0044,
+      "reward": 1.30078125,
+      "reward_std": 0.2698420286178589,
+      "rewards/accuracy_reward": 0.31640625,
+      "rewards/format_reward": 0.984375,
+      "step": 103
+    },
+    {
+      "completion_length": 105.11328125,
+      "epoch": 0.4315352697095436,
+      "grad_norm": 5.773576507821817,
+      "kl": 0.08251953125,
+      "learning_rate": 5.684647302904564e-07,
+      "loss": 0.0033,
+      "reward": 1.2421875,
+      "reward_std": 0.2864542603492737,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.9765625,
+      "step": 104
+    },
+    {
+      "completion_length": 87.703125,
+      "epoch": 0.43568464730290457,
+      "grad_norm": 11.513429520019525,
+      "kl": 0.109375,
+      "learning_rate": 5.643153526970955e-07,
+      "loss": 0.0044,
+      "reward": 1.27734375,
+      "reward_std": 0.2987739145755768,
+      "rewards/accuracy_reward": 0.2890625,
+      "rewards/format_reward": 0.98828125,
+      "step": 105
+    },
+    {
+      "completion_length": 85.6328125,
+      "epoch": 0.43983402489626555,
+      "grad_norm": 7.2230530225685134,
+      "kl": 0.10693359375,
+      "learning_rate": 5.601659751037344e-07,
+      "loss": 0.0043,
+      "reward": 1.203125,
+      "reward_std": 0.2508074939250946,
+      "rewards/accuracy_reward": 0.2109375,
+      "rewards/format_reward": 0.9921875,
+      "step": 106
+    },
+    {
+      "completion_length": 91.26171875,
+      "epoch": 0.44398340248962653,
+      "grad_norm": 11.723926666762416,
+      "kl": 0.0849609375,
+      "learning_rate": 5.560165975103735e-07,
+      "loss": 0.0034,
+      "reward": 1.2578125,
+      "reward_std": 0.24281391501426697,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.984375,
+      "step": 107
+    },
+    {
+      "completion_length": 85.66796875,
+      "epoch": 0.44813278008298757,
+      "grad_norm": 19.25236764754529,
+      "kl": 0.09716796875,
+      "learning_rate": 5.518672199170124e-07,
+      "loss": 0.0039,
+      "reward": 1.29296875,
+      "reward_std": 0.27077654004096985,
+      "rewards/accuracy_reward": 0.29296875,
+      "rewards/format_reward": 1.0,
+      "step": 108
+    },
+    {
+      "completion_length": 95.21875,
+      "epoch": 0.45228215767634855,
+      "grad_norm": 3.839508040726909,
+      "kl": 0.1015625,
+      "learning_rate": 5.477178423236515e-07,
+      "loss": 0.0041,
+      "reward": 1.171875,
+      "reward_std": 0.26314786076545715,
+      "rewards/accuracy_reward": 0.203125,
+      "rewards/format_reward": 0.96875,
+      "step": 109
+    },
+    {
+      "completion_length": 87.390625,
+      "epoch": 0.45643153526970953,
+      "grad_norm": 4.757921717516325,
+      "kl": 0.1025390625,
+      "learning_rate": 5.435684647302903e-07,
+      "loss": 0.0041,
+      "reward": 1.234375,
+      "reward_std": 0.249478280544281,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.9921875,
+      "step": 110
+    },
+    {
+      "completion_length": 90.4609375,
+      "epoch": 0.4605809128630705,
+      "grad_norm": 12.145850813087781,
+      "kl": 0.08984375,
+      "learning_rate": 5.394190871369294e-07,
+      "loss": 0.0036,
+      "reward": 1.1640625,
+      "reward_std": 0.17517516016960144,
+      "rewards/accuracy_reward": 0.16796875,
+      "rewards/format_reward": 0.99609375,
+      "step": 111
+    },
+    {
+      "completion_length": 94.23046875,
+      "epoch": 0.46473029045643155,
+      "grad_norm": 11.118828549631772,
+      "kl": 0.08740234375,
+      "learning_rate": 5.352697095435685e-07,
+      "loss": 0.0035,
+      "reward": 1.203125,
+      "reward_std": 0.30724799633026123,
+      "rewards/accuracy_reward": 0.22265625,
+      "rewards/format_reward": 0.98046875,
+      "step": 112
+    },
+    {
+      "completion_length": 98.54296875,
+      "epoch": 0.46887966804979253,
+      "grad_norm": 12.335705990530665,
+      "kl": 0.091796875,
+      "learning_rate": 5.311203319502074e-07,
+      "loss": 0.0037,
+      "reward": 1.3515625,
+      "reward_std": 0.2899443805217743,
+      "rewards/accuracy_reward": 0.3671875,
+      "rewards/format_reward": 0.984375,
+      "step": 113
+    },
+    {
+      "completion_length": 87.42578125,
+      "epoch": 0.4730290456431535,
+      "grad_norm": 4.2994268361338,
+      "kl": 0.10498046875,
+      "learning_rate": 5.269709543568465e-07,
+      "loss": 0.0042,
+      "reward": 1.171875,
+      "reward_std": 0.22647252678871155,
+      "rewards/accuracy_reward": 0.18359375,
+      "rewards/format_reward": 0.98828125,
+      "step": 114
+    },
+    {
+      "completion_length": 90.40234375,
+      "epoch": 0.47717842323651455,
+      "grad_norm": 7.3486884231249805,
+      "kl": 0.10205078125,
+      "learning_rate": 5.228215767634854e-07,
+      "loss": 0.0041,
+      "reward": 1.1796875,
+      "reward_std": 0.26744821667671204,
+      "rewards/accuracy_reward": 0.19140625,
+      "rewards/format_reward": 0.98828125,
+      "step": 115
+    },
+    {
+      "completion_length": 93.9921875,
+      "epoch": 0.48132780082987553,
+      "grad_norm": 21.449341096677863,
+      "kl": 0.09765625,
+      "learning_rate": 5.186721991701245e-07,
+      "loss": 0.0039,
+      "reward": 1.27734375,
+      "reward_std": 0.3352429270744324,
+      "rewards/accuracy_reward": 0.296875,
+      "rewards/format_reward": 0.98046875,
+      "step": 116
+    },
+    {
+      "completion_length": 97.37890625,
+      "epoch": 0.4854771784232365,
+      "grad_norm": 3.291074287652009,
+      "kl": 0.0830078125,
+      "learning_rate": 5.145228215767634e-07,
+      "loss": 0.0033,
+      "reward": 1.234375,
+      "reward_std": 0.2131577730178833,
+      "rewards/accuracy_reward": 0.25390625,
+      "rewards/format_reward": 0.98046875,
+      "step": 117
+    },
+    {
+      "completion_length": 89.75390625,
+      "epoch": 0.4896265560165975,
+      "grad_norm": 8.052246239154032,
+      "kl": 0.11083984375,
+      "learning_rate": 5.103734439834024e-07,
+      "loss": 0.0044,
+      "reward": 1.140625,
+      "reward_std": 0.19178494811058044,
+      "rewards/accuracy_reward": 0.16796875,
+      "rewards/format_reward": 0.97265625,
+      "step": 118
+    },
+    {
+      "completion_length": 91.37890625,
+      "epoch": 0.49377593360995853,
+      "grad_norm": 20.497960348017255,
+      "kl": 0.11865234375,
+      "learning_rate": 5.062240663900415e-07,
+      "loss": 0.0047,
+      "reward": 1.2421875,
+      "reward_std": 0.3268205523490906,
+      "rewards/accuracy_reward": 0.25390625,
+      "rewards/format_reward": 0.98828125,
+      "step": 119
+    },
+    {
+      "completion_length": 109.734375,
+      "epoch": 0.4979253112033195,
+      "grad_norm": 5.486079926596014,
+      "kl": 0.07568359375,
+      "learning_rate": 5.020746887966804e-07,
+      "loss": 0.003,
+      "reward": 1.125,
+      "reward_std": 0.2778772711753845,
+      "rewards/accuracy_reward": 0.1875,
+      "rewards/format_reward": 0.9375,
+      "step": 120
+    },
+    {
+      "completion_length": 91.80078125,
+      "epoch": 0.5020746887966805,
+      "grad_norm": 2.85121433781658,
+      "kl": 0.09130859375,
+      "learning_rate": 4.979253112033195e-07,
+      "loss": 0.0036,
+      "reward": 1.3359375,
+      "reward_std": 0.21765941381454468,
+      "rewards/accuracy_reward": 0.33984375,
+      "rewards/format_reward": 0.99609375,
+      "step": 121
+    },
+    {
+      "completion_length": 108.34765625,
+      "epoch": 0.5062240663900415,
+      "grad_norm": 2.5487467805188415,
+      "kl": 0.0703125,
+      "learning_rate": 4.937759336099585e-07,
+      "loss": 0.0028,
+      "reward": 1.20703125,
+      "reward_std": 0.24190281331539154,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 0.97265625,
+      "step": 122
+    },
+    {
+      "completion_length": 88.4296875,
+      "epoch": 0.5103734439834025,
+      "grad_norm": 9.052955326669567,
+      "kl": 0.09326171875,
+      "learning_rate": 4.896265560165974e-07,
+      "loss": 0.0037,
+      "reward": 1.30078125,
+      "reward_std": 0.2738521099090576,
+      "rewards/accuracy_reward": 0.3203125,
+      "rewards/format_reward": 0.98046875,
+      "step": 123
+    },
+    {
+      "completion_length": 92.03125,
+      "epoch": 0.5145228215767634,
+      "grad_norm": 4.78722259192626,
+      "kl": 0.10595703125,
+      "learning_rate": 4.854771784232365e-07,
+      "loss": 0.0042,
+      "reward": 1.25,
+      "reward_std": 0.2780294418334961,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.96484375,
+      "step": 124
+    },
+    {
+      "completion_length": 90.37109375,
+      "epoch": 0.5186721991701245,
+      "grad_norm": 31.7592456228116,
+      "kl": 0.1328125,
+      "learning_rate": 4.813278008298756e-07,
+      "loss": 0.0053,
+      "reward": 1.26171875,
+      "reward_std": 0.264590859413147,
+      "rewards/accuracy_reward": 0.296875,
+      "rewards/format_reward": 0.96484375,
+      "step": 125
+    },
+    {
+      "completion_length": 92.3203125,
+      "epoch": 0.5228215767634855,
+      "grad_norm": 3.431661013903984,
+      "kl": 0.095703125,
+      "learning_rate": 4.771784232365145e-07,
+      "loss": 0.0038,
+      "reward": 1.2890625,
+      "reward_std": 0.2519586980342865,
+      "rewards/accuracy_reward": 0.296875,
+      "rewards/format_reward": 0.9921875,
+      "step": 126
+    },
+    {
+      "completion_length": 93.17578125,
+      "epoch": 0.5269709543568465,
+      "grad_norm": 4.2013084848725555,
+      "kl": 0.091796875,
+      "learning_rate": 4.7302904564315346e-07,
+      "loss": 0.0037,
+      "reward": 1.24609375,
+      "reward_std": 0.22338411211967468,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.98046875,
+      "step": 127
+    },
+    {
+      "completion_length": 98.3359375,
+      "epoch": 0.5311203319502075,
+      "grad_norm": 10.80790694765271,
+      "kl": 0.087890625,
+      "learning_rate": 4.6887966804979253e-07,
+      "loss": 0.0035,
+      "reward": 1.171875,
+      "reward_std": 0.2016046792268753,
+      "rewards/accuracy_reward": 0.18359375,
+      "rewards/format_reward": 0.98828125,
+      "step": 128
+    },
+    {
+      "completion_length": 102.2890625,
+      "epoch": 0.5352697095435685,
+      "grad_norm": 2.602202807896544,
+      "kl": 0.08251953125,
+      "learning_rate": 4.6473029045643154e-07,
+      "loss": 0.0033,
+      "reward": 1.2109375,
+      "reward_std": 0.25826555490493774,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.96484375,
+      "step": 129
+    },
+    {
+      "completion_length": 96.62109375,
+      "epoch": 0.5394190871369294,
+      "grad_norm": 9.873703703904674,
+      "kl": 0.10888671875,
+      "learning_rate": 4.605809128630705e-07,
+      "loss": 0.0044,
+      "reward": 1.22265625,
+      "reward_std": 0.26187431812286377,
+      "rewards/accuracy_reward": 0.25390625,
+      "rewards/format_reward": 0.96875,
+      "step": 130
+    },
+    {
+      "completion_length": 98.99609375,
+      "epoch": 0.5435684647302904,
+      "grad_norm": 3.67428920521419,
+      "kl": 0.0927734375,
+      "learning_rate": 4.564315352697095e-07,
+      "loss": 0.0037,
+      "reward": 1.16796875,
+      "reward_std": 0.2292679101228714,
+      "rewards/accuracy_reward": 0.1953125,
+      "rewards/format_reward": 0.97265625,
+      "step": 131
+    },
+    {
+      "completion_length": 92.67578125,
+      "epoch": 0.5477178423236515,
+      "grad_norm": 4.679154410147011,
+      "kl": 0.08544921875,
+      "learning_rate": 4.522821576763485e-07,
+      "loss": 0.0034,
+      "reward": 1.3359375,
+      "reward_std": 0.35772189497947693,
+      "rewards/accuracy_reward": 0.35546875,
+      "rewards/format_reward": 0.98046875,
+      "step": 132
+    },
+    {
+      "completion_length": 106.16015625,
+      "epoch": 0.5518672199170125,
+      "grad_norm": 6.527275895854011,
+      "kl": 0.07763671875,
+      "learning_rate": 4.481327800829875e-07,
+      "loss": 0.0031,
+      "reward": 1.19140625,
+      "reward_std": 0.2933124899864197,
+      "rewards/accuracy_reward": 0.23046875,
+      "rewards/format_reward": 0.9609375,
+      "step": 133
+    },
+    {
+      "completion_length": 98.29296875,
+      "epoch": 0.5560165975103735,
+      "grad_norm": 6.290852562410683,
+      "kl": 0.08349609375,
+      "learning_rate": 4.4398340248962654e-07,
+      "loss": 0.0033,
+      "reward": 1.24609375,
+      "reward_std": 0.27056628465652466,
+      "rewards/accuracy_reward": 0.27734375,
+      "rewards/format_reward": 0.96875,
+      "step": 134
+    },
+    {
+      "completion_length": 103.40625,
+      "epoch": 0.5601659751037344,
+      "grad_norm": 9.043101857772628,
+      "kl": 0.083984375,
+      "learning_rate": 4.3983402489626555e-07,
+      "loss": 0.0034,
+      "reward": 1.3046875,
+      "reward_std": 0.2771390676498413,
+      "rewards/accuracy_reward": 0.32421875,
+      "rewards/format_reward": 0.98046875,
+      "step": 135
+    },
+    {
+      "completion_length": 103.0078125,
+      "epoch": 0.5643153526970954,
+      "grad_norm": 4.096567104156379,
+      "kl": 0.08349609375,
+      "learning_rate": 4.3568464730290456e-07,
+      "loss": 0.0033,
+      "reward": 1.20703125,
+      "reward_std": 0.23877452313899994,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.9609375,
+      "step": 136
+    },
+    {
+      "completion_length": 98.453125,
+      "epoch": 0.5684647302904564,
+      "grad_norm": 5.621101251975948,
+      "kl": 0.1015625,
+      "learning_rate": 4.3153526970954357e-07,
+      "loss": 0.0041,
+      "reward": 1.19921875,
+      "reward_std": 0.27381569147109985,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.94921875,
+      "step": 137
+    },
+    {
+      "completion_length": 110.04296875,
+      "epoch": 0.5726141078838174,
+      "grad_norm": 8.904930677056841,
+      "kl": 0.0712890625,
+      "learning_rate": 4.273858921161825e-07,
+      "loss": 0.0029,
+      "reward": 1.2109375,
+      "reward_std": 0.3127846121788025,
+      "rewards/accuracy_reward": 0.2578125,
+      "rewards/format_reward": 0.953125,
+      "step": 138
+    },
+    {
+      "completion_length": 107.1875,
+      "epoch": 0.5767634854771784,
+      "grad_norm": 4.817083135230505,
+      "kl": 0.07666015625,
+      "learning_rate": 4.2323651452282153e-07,
+      "loss": 0.0031,
+      "reward": 1.16796875,
+      "reward_std": 0.25475409626960754,
+      "rewards/accuracy_reward": 0.19921875,
+      "rewards/format_reward": 0.96875,
+      "step": 139
+    },
+    {
+      "completion_length": 105.25,
+      "epoch": 0.5809128630705395,
+      "grad_norm": 29.934647705971432,
+      "kl": 0.3984375,
+      "learning_rate": 4.190871369294606e-07,
+      "loss": 0.016,
+      "reward": 1.30859375,
+      "reward_std": 0.35114049911499023,
+      "rewards/accuracy_reward": 0.33984375,
+      "rewards/format_reward": 0.96875,
+      "step": 140
+    },
+    {
+      "completion_length": 100.92578125,
+      "epoch": 0.5850622406639004,
+      "grad_norm": 2.5340527761032625,
+      "kl": 0.0830078125,
+      "learning_rate": 4.149377593360996e-07,
+      "loss": 0.0033,
+      "reward": 1.2109375,
+      "reward_std": 0.2652069628238678,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 0.9765625,
+      "step": 141
+    },
+    {
+      "completion_length": 101.79296875,
+      "epoch": 0.5892116182572614,
+      "grad_norm": 3.363293811701697,
+      "kl": 0.08984375,
+      "learning_rate": 4.1078838174273857e-07,
+      "loss": 0.0036,
+      "reward": 1.2421875,
+      "reward_std": 0.262579083442688,
+      "rewards/accuracy_reward": 0.28125,
+      "rewards/format_reward": 0.9609375,
+      "step": 142
+    },
+    {
+      "completion_length": 98.31640625,
+      "epoch": 0.5933609958506224,
+      "grad_norm": 3.049695551614608,
+      "kl": 0.091796875,
+      "learning_rate": 4.066390041493776e-07,
+      "loss": 0.0037,
+      "reward": 1.19140625,
+      "reward_std": 0.2985348701477051,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 0.95703125,
+      "step": 143
+    },
+    {
+      "completion_length": 97.63671875,
+      "epoch": 0.5975103734439834,
+      "grad_norm": 4.921448840015921,
+      "kl": 0.08349609375,
+      "learning_rate": 4.024896265560166e-07,
+      "loss": 0.0033,
+      "reward": 1.33984375,
+      "reward_std": 0.27990859746932983,
+      "rewards/accuracy_reward": 0.3515625,
+      "rewards/format_reward": 0.98828125,
+      "step": 144
+    },
+    {
+      "completion_length": 119.609375,
+      "epoch": 0.6016597510373444,
+      "grad_norm": 2.821821714985559,
+      "kl": 0.06884765625,
+      "learning_rate": 3.9834024896265554e-07,
+      "loss": 0.0027,
+      "reward": 1.2578125,
+      "reward_std": 0.3430861532688141,
+      "rewards/accuracy_reward": 0.30078125,
+      "rewards/format_reward": 0.95703125,
+      "step": 145
+    },
+    {
+      "completion_length": 100.03125,
+      "epoch": 0.6058091286307054,
+      "grad_norm": 2.9588689744644654,
+      "kl": 0.103515625,
+      "learning_rate": 3.9419087136929455e-07,
+      "loss": 0.0041,
+      "reward": 1.20703125,
+      "reward_std": 0.35439443588256836,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.93359375,
+      "step": 146
+    },
+    {
+      "completion_length": 106.19140625,
+      "epoch": 0.6099585062240664,
+      "grad_norm": 5.79280616422922,
+      "kl": 0.125,
+      "learning_rate": 3.900414937759336e-07,
+      "loss": 0.005,
+      "reward": 1.203125,
+      "reward_std": 0.29194867610931396,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.9609375,
+      "step": 147
+    },
+    {
+      "completion_length": 95.98828125,
+      "epoch": 0.6141078838174274,
+      "grad_norm": 5.4671468471721605,
+      "kl": 0.10986328125,
+      "learning_rate": 3.8589211618257263e-07,
+      "loss": 0.0044,
+      "reward": 1.22265625,
+      "reward_std": 0.2797732949256897,
+      "rewards/accuracy_reward": 0.23828125,
+      "rewards/format_reward": 0.984375,
+      "step": 148
+    },
+    {
+      "completion_length": 88.03515625,
+      "epoch": 0.6182572614107884,
+      "grad_norm": 8.553849856772748,
+      "kl": 0.11376953125,
+      "learning_rate": 3.817427385892116e-07,
+      "loss": 0.0045,
+      "reward": 1.265625,
+      "reward_std": 0.30274513363838196,
+      "rewards/accuracy_reward": 0.30078125,
+      "rewards/format_reward": 0.96484375,
+      "step": 149
+    },
+    {
+      "completion_length": 107.9140625,
+      "epoch": 0.6224066390041494,
+      "grad_norm": 4.437784186558576,
+      "kl": 0.08349609375,
+      "learning_rate": 3.775933609958506e-07,
+      "loss": 0.0033,
+      "reward": 1.2890625,
+      "reward_std": 0.3589634895324707,
+      "rewards/accuracy_reward": 0.3203125,
+      "rewards/format_reward": 0.96875,
+      "step": 150
+    },
+    {
+      "completion_length": 91.25,
+      "epoch": 0.6265560165975104,
+      "grad_norm": 7.218676225343139,
+      "kl": 0.09130859375,
+      "learning_rate": 3.734439834024896e-07,
+      "loss": 0.0037,
+      "reward": 1.2265625,
+      "reward_std": 0.2923767566680908,
+      "rewards/accuracy_reward": 0.25390625,
+      "rewards/format_reward": 0.97265625,
+      "step": 151
+    },
+    {
+      "completion_length": 102.1796875,
+      "epoch": 0.6307053941908713,
+      "grad_norm": 5.625873375289916,
+      "kl": 0.08251953125,
+      "learning_rate": 3.692946058091286e-07,
+      "loss": 0.0033,
+      "reward": 1.26953125,
+      "reward_std": 0.28951701521873474,
+      "rewards/accuracy_reward": 0.296875,
+      "rewards/format_reward": 0.97265625,
+      "step": 152
+    },
+    {
+      "completion_length": 94.6875,
+      "epoch": 0.6348547717842323,
+      "grad_norm": 3.3046695270707143,
+      "kl": 0.08740234375,
+      "learning_rate": 3.6514522821576757e-07,
+      "loss": 0.0035,
+      "reward": 1.20703125,
+      "reward_std": 0.3152885437011719,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.96484375,
+      "step": 153
+    },
+    {
+      "completion_length": 93.8359375,
+      "epoch": 0.6390041493775933,
+      "grad_norm": 3.2380098166032174,
+      "kl": 0.0927734375,
+      "learning_rate": 3.6099585062240664e-07,
+      "loss": 0.0037,
+      "reward": 1.41015625,
+      "reward_std": 0.3835912346839905,
+      "rewards/accuracy_reward": 0.421875,
+      "rewards/format_reward": 0.98828125,
+      "step": 154
+    },
+    {
+      "completion_length": 100.3828125,
+      "epoch": 0.6431535269709544,
+      "grad_norm": 3.3134631461874555,
+      "kl": 0.091796875,
+      "learning_rate": 3.5684647302904565e-07,
+      "loss": 0.0037,
+      "reward": 1.21875,
+      "reward_std": 0.36876800656318665,
+      "rewards/accuracy_reward": 0.25390625,
+      "rewards/format_reward": 0.96484375,
+      "step": 155
+    },
+    {
+      "completion_length": 85.98828125,
+      "epoch": 0.6473029045643154,
+      "grad_norm": 8.891623678352106,
+      "kl": 0.1044921875,
+      "learning_rate": 3.5269709543568466e-07,
+      "loss": 0.0042,
+      "reward": 1.34375,
+      "reward_std": 0.2681604027748108,
+      "rewards/accuracy_reward": 0.3515625,
+      "rewards/format_reward": 0.9921875,
+      "step": 156
+    },
+    {
+      "completion_length": 96.3125,
+      "epoch": 0.6514522821576764,
+      "grad_norm": 10.997826794239984,
+      "kl": 0.103515625,
+      "learning_rate": 3.485477178423236e-07,
+      "loss": 0.0041,
+      "reward": 1.30859375,
+      "reward_std": 0.30442675948143005,
+      "rewards/accuracy_reward": 0.33203125,
+      "rewards/format_reward": 0.9765625,
+      "step": 157
+    },
+    {
+      "completion_length": 97.25,
+      "epoch": 0.6556016597510373,
+      "grad_norm": 15.501954515999833,
+      "kl": 0.11181640625,
+      "learning_rate": 3.443983402489626e-07,
+      "loss": 0.0045,
+      "reward": 1.24609375,
+      "reward_std": 0.3243793845176697,
+      "rewards/accuracy_reward": 0.2890625,
+      "rewards/format_reward": 0.95703125,
+      "step": 158
+    },
+    {
+      "completion_length": 85.5546875,
+      "epoch": 0.6597510373443983,
+      "grad_norm": 15.699833790155175,
+      "kl": 0.125,
+      "learning_rate": 3.4024896265560163e-07,
+      "loss": 0.005,
+      "reward": 1.2421875,
+      "reward_std": 0.2799358367919922,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.9765625,
+      "step": 159
+    },
+    {
+      "completion_length": 87.35546875,
+      "epoch": 0.6639004149377593,
+      "grad_norm": 21.010504513986813,
+      "kl": 0.09423828125,
+      "learning_rate": 3.360995850622407e-07,
+      "loss": 0.0038,
+      "reward": 1.1953125,
+      "reward_std": 0.19621142745018005,
+      "rewards/accuracy_reward": 0.203125,
+      "rewards/format_reward": 0.9921875,
+      "step": 160
+    },
+    {
+      "completion_length": 79.609375,
+      "epoch": 0.6680497925311203,
+      "grad_norm": 15.76259223021215,
+      "kl": 0.12890625,
+      "learning_rate": 3.3195020746887966e-07,
+      "loss": 0.0052,
+      "reward": 1.28515625,
+      "reward_std": 0.2734532654285431,
+      "rewards/accuracy_reward": 0.30859375,
+      "rewards/format_reward": 0.9765625,
+      "step": 161
+    },
+    {
+      "completion_length": 92.03515625,
+      "epoch": 0.6721991701244814,
+      "grad_norm": 10.404250366698443,
+      "kl": 0.11865234375,
+      "learning_rate": 3.2780082987551867e-07,
+      "loss": 0.0047,
+      "reward": 1.12890625,
+      "reward_std": 0.13782215118408203,
+      "rewards/accuracy_reward": 0.14453125,
+      "rewards/format_reward": 0.984375,
+      "step": 162
+    },
+    {
+      "completion_length": 87.2421875,
+      "epoch": 0.6763485477178424,
+      "grad_norm": 14.551252603892712,
+      "kl": 0.09912109375,
+      "learning_rate": 3.236514522821577e-07,
+      "loss": 0.004,
+      "reward": 1.2421875,
+      "reward_std": 0.27508556842803955,
+      "rewards/accuracy_reward": 0.2578125,
+      "rewards/format_reward": 0.984375,
+      "step": 163
+    },
+    {
+      "completion_length": 86.05078125,
+      "epoch": 0.6804979253112033,
+      "grad_norm": 19.037085273013254,
+      "kl": 0.1513671875,
+      "learning_rate": 3.1950207468879663e-07,
+      "loss": 0.006,
+      "reward": 1.39453125,
+      "reward_std": 0.3236626982688904,
+      "rewards/accuracy_reward": 0.40234375,
+      "rewards/format_reward": 0.9921875,
+      "step": 164
+    },
+    {
+      "completion_length": 85.9609375,
+      "epoch": 0.6846473029045643,
+      "grad_norm": 43.15683291911658,
+      "kl": 0.1416015625,
+      "learning_rate": 3.1535269709543564e-07,
+      "loss": 0.0057,
+      "reward": 1.30859375,
+      "reward_std": 0.2828989326953888,
+      "rewards/accuracy_reward": 0.33984375,
+      "rewards/format_reward": 0.96875,
+      "step": 165
+    },
+    {
+      "completion_length": 88.8671875,
+      "epoch": 0.6887966804979253,
+      "grad_norm": 10.572313497625037,
+      "kl": 0.0927734375,
+      "learning_rate": 3.1120331950207465e-07,
+      "loss": 0.0037,
+      "reward": 1.34375,
+      "reward_std": 0.2745204567909241,
+      "rewards/accuracy_reward": 0.359375,
+      "rewards/format_reward": 0.984375,
+      "step": 166
+    },
+    {
+      "completion_length": 93.8125,
+      "epoch": 0.6929460580912863,
+      "grad_norm": 5.703362063681082,
+      "kl": 0.103515625,
+      "learning_rate": 3.070539419087137e-07,
+      "loss": 0.0041,
+      "reward": 1.1796875,
+      "reward_std": 0.24057602882385254,
+      "rewards/accuracy_reward": 0.1953125,
+      "rewards/format_reward": 0.984375,
+      "step": 167
+    },
+    {
+      "completion_length": 88.37109375,
+      "epoch": 0.6970954356846473,
+      "grad_norm": 26054.768735426984,
+      "kl": 3.203125,
+      "learning_rate": 3.029045643153527e-07,
+      "loss": 0.1277,
+      "reward": 1.234375,
+      "reward_std": 0.22063545882701874,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.984375,
+      "step": 168
+    },
+    {
+      "completion_length": 94.67578125,
+      "epoch": 0.7012448132780082,
+      "grad_norm": 15.618598784917781,
+      "kl": 0.10400390625,
+      "learning_rate": 2.987551867219917e-07,
+      "loss": 0.0042,
+      "reward": 1.2890625,
+      "reward_std": 0.27328526973724365,
+      "rewards/accuracy_reward": 0.30078125,
+      "rewards/format_reward": 0.98828125,
+      "step": 169
+    },
+    {
+      "completion_length": 99.70703125,
+      "epoch": 0.7053941908713693,
+      "grad_norm": 9.871899460079366,
+      "kl": 0.09521484375,
+      "learning_rate": 2.946058091286307e-07,
+      "loss": 0.0038,
+      "reward": 1.3671875,
+      "reward_std": 0.3782796263694763,
+      "rewards/accuracy_reward": 0.38671875,
+      "rewards/format_reward": 0.98046875,
+      "step": 170
+    },
+    {
+      "completion_length": 80.71875,
+      "epoch": 0.7095435684647303,
+      "grad_norm": 18.124016004514953,
+      "kl": 0.10546875,
+      "learning_rate": 2.904564315352697e-07,
+      "loss": 0.0042,
+      "reward": 1.21875,
+      "reward_std": 0.22859051823616028,
+      "rewards/accuracy_reward": 0.2265625,
+      "rewards/format_reward": 0.9921875,
+      "step": 171
+    },
+    {
+      "completion_length": 95.65625,
+      "epoch": 0.7136929460580913,
+      "grad_norm": 11.410453654872125,
+      "kl": 0.08837890625,
+      "learning_rate": 2.8630705394190866e-07,
+      "loss": 0.0035,
+      "reward": 1.265625,
+      "reward_std": 0.3315829634666443,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.98046875,
+      "step": 172
+    },
+    {
+      "completion_length": 84.04296875,
+      "epoch": 0.7178423236514523,
+      "grad_norm": 11.679683395107324,
+      "kl": 0.1162109375,
+      "learning_rate": 2.821576763485477e-07,
+      "loss": 0.0046,
+      "reward": 1.2578125,
+      "reward_std": 0.2511623501777649,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.9921875,
+      "step": 173
+    },
+    {
+      "completion_length": 86.625,
+      "epoch": 0.7219917012448133,
+      "grad_norm": 12.07785638477025,
+      "kl": 0.11083984375,
+      "learning_rate": 2.7800829875518674e-07,
+      "loss": 0.0044,
+      "reward": 1.22265625,
+      "reward_std": 0.17005667090415955,
+      "rewards/accuracy_reward": 0.23828125,
+      "rewards/format_reward": 0.984375,
+      "step": 174
+    },
+    {
+      "completion_length": 105.43359375,
+      "epoch": 0.7261410788381742,
+      "grad_norm": 8.597610653971167,
+      "kl": 0.08984375,
+      "learning_rate": 2.7385892116182575e-07,
+      "loss": 0.0036,
+      "reward": 1.29296875,
+      "reward_std": 0.2580205202102661,
+      "rewards/accuracy_reward": 0.3359375,
+      "rewards/format_reward": 0.95703125,
+      "step": 175
+    },
+    {
+      "completion_length": 91.1171875,
+      "epoch": 0.7302904564315352,
+      "grad_norm": 10.411659860843175,
+      "kl": 0.134765625,
+      "learning_rate": 2.697095435684647e-07,
+      "loss": 0.0054,
+      "reward": 1.28125,
+      "reward_std": 0.33533939719200134,
+      "rewards/accuracy_reward": 0.30859375,
+      "rewards/format_reward": 0.97265625,
+      "step": 176
+    },
+    {
+      "completion_length": 84.40625,
+      "epoch": 0.7344398340248963,
+      "grad_norm": 8.502997007157745,
+      "kl": 0.09716796875,
+      "learning_rate": 2.655601659751037e-07,
+      "loss": 0.0039,
+      "reward": 1.3203125,
+      "reward_std": 0.303774893283844,
+      "rewards/accuracy_reward": 0.3203125,
+      "rewards/format_reward": 1.0,
+      "step": 177
+    },
+    {
+      "completion_length": 80.4296875,
+      "epoch": 0.7385892116182573,
+      "grad_norm": 10.088884677713335,
+      "kl": 0.111328125,
+      "learning_rate": 2.614107883817427e-07,
+      "loss": 0.0045,
+      "reward": 1.24609375,
+      "reward_std": 0.27983221411705017,
+      "rewards/accuracy_reward": 0.25390625,
+      "rewards/format_reward": 0.9921875,
+      "step": 178
+    },
+    {
+      "completion_length": 102.3359375,
+      "epoch": 0.7427385892116183,
+      "grad_norm": 7.955033668582297,
+      "kl": 0.0859375,
+      "learning_rate": 2.572614107883817e-07,
+      "loss": 0.0034,
+      "reward": 1.2578125,
+      "reward_std": 0.1971781998872757,
+      "rewards/accuracy_reward": 0.26953125,
+      "rewards/format_reward": 0.98828125,
+      "step": 179
+    },
+    {
+      "completion_length": 91.73046875,
+      "epoch": 0.7468879668049793,
+      "grad_norm": 3.07504994450314,
+      "kl": 0.09716796875,
+      "learning_rate": 2.5311203319502074e-07,
+      "loss": 0.0039,
+      "reward": 1.31640625,
+      "reward_std": 0.23474639654159546,
+      "rewards/accuracy_reward": 0.328125,
+      "rewards/format_reward": 0.98828125,
+      "step": 180
+    },
+    {
+      "completion_length": 89.3984375,
+      "epoch": 0.7510373443983402,
+      "grad_norm": 5.922630830808792,
+      "kl": 0.12060546875,
+      "learning_rate": 2.4896265560165975e-07,
+      "loss": 0.0048,
+      "reward": 1.3828125,
+      "reward_std": 0.2881060838699341,
+      "rewards/accuracy_reward": 0.390625,
+      "rewards/format_reward": 0.9921875,
+      "step": 181
+    },
+    {
+      "completion_length": 98.4296875,
+      "epoch": 0.7551867219917012,
+      "grad_norm": 5.611561605457396,
+      "kl": 0.099609375,
+      "learning_rate": 2.448132780082987e-07,
+      "loss": 0.004,
+      "reward": 1.1953125,
+      "reward_std": 0.26515352725982666,
+      "rewards/accuracy_reward": 0.21875,
+      "rewards/format_reward": 0.9765625,
+      "step": 182
+    },
+    {
+      "completion_length": 95.12890625,
+      "epoch": 0.7593360995850622,
+      "grad_norm": 3.07356690502102,
+      "kl": 0.1044921875,
+      "learning_rate": 2.406639004149378e-07,
+      "loss": 0.0042,
+      "reward": 1.1640625,
+      "reward_std": 0.23651200532913208,
+      "rewards/accuracy_reward": 0.1953125,
+      "rewards/format_reward": 0.96875,
+      "step": 183
+    },
+    {
+      "completion_length": 89.78125,
+      "epoch": 0.7634854771784232,
+      "grad_norm": 5.291830044415155,
+      "kl": 0.107421875,
+      "learning_rate": 2.3651452282157673e-07,
+      "loss": 0.0043,
+      "reward": 1.3046875,
+      "reward_std": 0.2729809880256653,
+      "rewards/accuracy_reward": 0.31640625,
+      "rewards/format_reward": 0.98828125,
+      "step": 184
+    },
+    {
+      "completion_length": 96.91796875,
+      "epoch": 0.7676348547717843,
+      "grad_norm": 6.323500099655262,
+      "kl": 0.0947265625,
+      "learning_rate": 2.3236514522821577e-07,
+      "loss": 0.0038,
+      "reward": 1.3125,
+      "reward_std": 0.29040926694869995,
+      "rewards/accuracy_reward": 0.3359375,
+      "rewards/format_reward": 0.9765625,
+      "step": 185
+    },
+    {
+      "completion_length": 85.296875,
+      "epoch": 0.7717842323651453,
+      "grad_norm": 38.28802098479183,
+      "kl": 0.11083984375,
+      "learning_rate": 2.2821576763485475e-07,
+      "loss": 0.0044,
+      "reward": 1.21875,
+      "reward_std": 0.2543177008628845,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.97265625,
+      "step": 186
+    },
+    {
+      "completion_length": 92.23046875,
+      "epoch": 0.7759336099585062,
+      "grad_norm": 3.3424642787888916,
+      "kl": 0.09375,
+      "learning_rate": 2.2406639004149376e-07,
+      "loss": 0.0037,
+      "reward": 1.30859375,
+      "reward_std": 0.2530517578125,
+      "rewards/accuracy_reward": 0.34765625,
+      "rewards/format_reward": 0.9609375,
+      "step": 187
+    },
+    {
+      "completion_length": 99.76171875,
+      "epoch": 0.7800829875518672,
+      "grad_norm": 4.333334591818117,
+      "kl": 0.0927734375,
+      "learning_rate": 2.1991701244813277e-07,
+      "loss": 0.0037,
+      "reward": 1.37109375,
+      "reward_std": 0.2535189092159271,
+      "rewards/accuracy_reward": 0.390625,
+      "rewards/format_reward": 0.98046875,
+      "step": 188
+    },
+    {
+      "completion_length": 94.41796875,
+      "epoch": 0.7842323651452282,
+      "grad_norm": 5.581784795886384,
+      "kl": 0.10693359375,
+      "learning_rate": 2.1576763485477178e-07,
+      "loss": 0.0043,
+      "reward": 1.23046875,
+      "reward_std": 0.27355122566223145,
+      "rewards/accuracy_reward": 0.26171875,
+      "rewards/format_reward": 0.96875,
+      "step": 189
+    },
+    {
+      "completion_length": 99.2109375,
+      "epoch": 0.7883817427385892,
+      "grad_norm": 5.627278869348173,
+      "kl": 0.09130859375,
+      "learning_rate": 2.1161825726141077e-07,
+      "loss": 0.0036,
+      "reward": 1.30859375,
+      "reward_std": 0.26744574308395386,
+      "rewards/accuracy_reward": 0.34765625,
+      "rewards/format_reward": 0.9609375,
+      "step": 190
+    },
+    {
+      "completion_length": 92.41796875,
+      "epoch": 0.7925311203319502,
+      "grad_norm": 5.131313447662913,
+      "kl": 0.095703125,
+      "learning_rate": 2.074688796680498e-07,
+      "loss": 0.0038,
+      "reward": 1.28515625,
+      "reward_std": 0.3070391118526459,
+      "rewards/accuracy_reward": 0.3046875,
+      "rewards/format_reward": 0.98046875,
+      "step": 191
+    },
+    {
+      "completion_length": 100.51953125,
+      "epoch": 0.7966804979253111,
+      "grad_norm": 4.031558423340292,
+      "kl": 0.0771484375,
+      "learning_rate": 2.033195020746888e-07,
+      "loss": 0.0031,
+      "reward": 1.32421875,
+      "reward_std": 0.220160573720932,
+      "rewards/accuracy_reward": 0.34765625,
+      "rewards/format_reward": 0.9765625,
+      "step": 192
+    },
+    {
+      "completion_length": 95.21484375,
+      "epoch": 0.8008298755186722,
+      "grad_norm": 7.853584345992108,
+      "kl": 0.11669921875,
+      "learning_rate": 1.9917012448132777e-07,
+      "loss": 0.0047,
+      "reward": 1.203125,
+      "reward_std": 0.2865522503852844,
+      "rewards/accuracy_reward": 0.234375,
+      "rewards/format_reward": 0.96875,
+      "step": 193
+    },
+    {
+      "completion_length": 101.75,
+      "epoch": 0.8049792531120332,
+      "grad_norm": 2.8908709714738614,
+      "kl": 0.08837890625,
+      "learning_rate": 1.950207468879668e-07,
+      "loss": 0.0035,
+      "reward": 1.234375,
+      "reward_std": 0.34296295046806335,
+      "rewards/accuracy_reward": 0.2890625,
+      "rewards/format_reward": 0.9453125,
+      "step": 194
+    },
+    {
+      "completion_length": 90.21875,
+      "epoch": 0.8091286307053942,
+      "grad_norm": 2.098330717758711,
+      "kl": 0.09228515625,
+      "learning_rate": 1.908713692946058e-07,
+      "loss": 0.0037,
+      "reward": 1.2265625,
+      "reward_std": 0.2507816553115845,
+      "rewards/accuracy_reward": 0.23828125,
+      "rewards/format_reward": 0.98828125,
+      "step": 195
+    },
+    {
+      "completion_length": 109.3828125,
+      "epoch": 0.8132780082987552,
+      "grad_norm": 2349.258527115438,
+      "kl": 6.90625,
+      "learning_rate": 1.867219917012448e-07,
+      "loss": 0.2761,
+      "reward": 1.21484375,
+      "reward_std": 0.264018177986145,
+      "rewards/accuracy_reward": 0.2421875,
+      "rewards/format_reward": 0.97265625,
+      "step": 196
+    },
+    {
+      "completion_length": 96.6640625,
+      "epoch": 0.8174273858921162,
+      "grad_norm": 16.535509899854667,
+      "kl": 0.1025390625,
+      "learning_rate": 1.8257261410788379e-07,
+      "loss": 0.0041,
+      "reward": 1.24609375,
+      "reward_std": 0.3564358055591583,
+      "rewards/accuracy_reward": 0.29296875,
+      "rewards/format_reward": 0.953125,
+      "step": 197
+    },
+    {
+      "completion_length": 88.40234375,
+      "epoch": 0.8215767634854771,
+      "grad_norm": 7.009787448907159,
+      "kl": 0.11181640625,
+      "learning_rate": 1.7842323651452282e-07,
+      "loss": 0.0045,
+      "reward": 1.37890625,
+      "reward_std": 0.3150339424610138,
+      "rewards/accuracy_reward": 0.41015625,
+      "rewards/format_reward": 0.96875,
+      "step": 198
+    },
+    {
+      "completion_length": 101.0,
+      "epoch": 0.8257261410788381,
+      "grad_norm": 1.4173394259733807,
+      "kl": 0.08544921875,
+      "learning_rate": 1.742738589211618e-07,
+      "loss": 0.0034,
+      "reward": 1.2265625,
+      "reward_std": 0.2616804838180542,
+      "rewards/accuracy_reward": 0.2578125,
+      "rewards/format_reward": 0.96875,
+      "step": 199
+    },
+    {
+      "completion_length": 87.8828125,
+      "epoch": 0.8298755186721992,
+      "grad_norm": 4.4140973894843825,
+      "kl": 0.09228515625,
+      "learning_rate": 1.7012448132780082e-07,
+      "loss": 0.0037,
+      "reward": 1.32421875,
+      "reward_std": 0.2538624703884125,
+      "rewards/accuracy_reward": 0.328125,
+      "rewards/format_reward": 0.99609375,
+      "step": 200
+    },
+    {
+      "completion_length": 99.31640625,
+      "epoch": 0.8340248962655602,
+      "grad_norm": 2.698781998865515,
+      "kl": 0.07275390625,
+      "learning_rate": 1.6597510373443983e-07,
+      "loss": 0.0029,
+      "reward": 1.2734375,
+      "reward_std": 0.3443642854690552,
+      "rewards/accuracy_reward": 0.3203125,
+      "rewards/format_reward": 0.953125,
+      "step": 201
+    },
+    {
+      "completion_length": 101.63671875,
+      "epoch": 0.8381742738589212,
+      "grad_norm": 3.34751147226265,
+      "kl": 0.09619140625,
+      "learning_rate": 1.6182572614107884e-07,
+      "loss": 0.0038,
+      "reward": 1.3203125,
+      "reward_std": 0.32434993982315063,
+      "rewards/accuracy_reward": 0.34375,
+      "rewards/format_reward": 0.9765625,
+      "step": 202
+    },
+    {
+      "completion_length": 95.63671875,
+      "epoch": 0.8423236514522822,
+      "grad_norm": 1.9689029980122599,
+      "kl": 0.08056640625,
+      "learning_rate": 1.5767634854771782e-07,
+      "loss": 0.0032,
+      "reward": 1.234375,
+      "reward_std": 0.2784101366996765,
+      "rewards/accuracy_reward": 0.25,
+      "rewards/format_reward": 0.984375,
+      "step": 203
+    },
+    {
+      "completion_length": 98.484375,
+      "epoch": 0.8464730290456431,
+      "grad_norm": 3.570750352423554,
+      "kl": 0.0830078125,
+      "learning_rate": 1.5352697095435686e-07,
+      "loss": 0.0033,
+      "reward": 1.3125,
+      "reward_std": 0.32842758297920227,
+      "rewards/accuracy_reward": 0.328125,
+      "rewards/format_reward": 0.984375,
+      "step": 204
+    },
+    {
+      "completion_length": 93.0390625,
+      "epoch": 0.8506224066390041,
+      "grad_norm": 3.372213441880041,
+      "kl": 0.076171875,
+      "learning_rate": 1.4937759336099584e-07,
+      "loss": 0.0031,
+      "reward": 1.24609375,
+      "reward_std": 0.2814468443393707,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.98046875,
+      "step": 205
+    },
+    {
+      "completion_length": 100.23046875,
+      "epoch": 0.8547717842323651,
+      "grad_norm": 3.861719509898381,
+      "kl": 0.09375,
+      "learning_rate": 1.4522821576763485e-07,
+      "loss": 0.0038,
+      "reward": 1.26953125,
+      "reward_std": 0.31658661365509033,
+      "rewards/accuracy_reward": 0.3125,
+      "rewards/format_reward": 0.95703125,
+      "step": 206
+    },
+    {
+      "completion_length": 89.0625,
+      "epoch": 0.8589211618257261,
+      "grad_norm": 6.172287251013444,
+      "kl": 0.087890625,
+      "learning_rate": 1.4107883817427386e-07,
+      "loss": 0.0035,
+      "reward": 1.3046875,
+      "reward_std": 0.25658145546913147,
+      "rewards/accuracy_reward": 0.31640625,
+      "rewards/format_reward": 0.98828125,
+      "step": 207
+    },
+    {
+      "completion_length": 98.75390625,
+      "epoch": 0.8630705394190872,
+      "grad_norm": 12.466091184278106,
+      "kl": 0.09521484375,
+      "learning_rate": 1.3692946058091287e-07,
+      "loss": 0.0038,
+      "reward": 1.1796875,
+      "reward_std": 0.28974291682243347,
+      "rewards/accuracy_reward": 0.21484375,
+      "rewards/format_reward": 0.96484375,
+      "step": 208
+    },
+    {
+      "completion_length": 85.42578125,
+      "epoch": 0.8672199170124482,
+      "grad_norm": 3.393938200300927,
+      "kl": 0.09033203125,
+      "learning_rate": 1.3278008298755186e-07,
+      "loss": 0.0036,
+      "reward": 1.41015625,
+      "reward_std": 0.26117590069770813,
+      "rewards/accuracy_reward": 0.4296875,
+      "rewards/format_reward": 0.98046875,
+      "step": 209
+    },
+    {
+      "completion_length": 106.23046875,
+      "epoch": 0.8713692946058091,
+      "grad_norm": 6.618744333693797,
+      "kl": 0.0732421875,
+      "learning_rate": 1.2863070539419084e-07,
+      "loss": 0.0029,
+      "reward": 1.27734375,
+      "reward_std": 0.3217734694480896,
+      "rewards/accuracy_reward": 0.29296875,
+      "rewards/format_reward": 0.984375,
+      "step": 210
+    },
+    {
+      "completion_length": 94.83984375,
+      "epoch": 0.8755186721991701,
+      "grad_norm": 3.712536977931616,
+      "kl": 0.0908203125,
+      "learning_rate": 1.2448132780082988e-07,
+      "loss": 0.0036,
+      "reward": 1.34375,
+      "reward_std": 0.30023884773254395,
+      "rewards/accuracy_reward": 0.36328125,
+      "rewards/format_reward": 0.98046875,
+      "step": 211
+    },
+    {
+      "completion_length": 101.30078125,
+      "epoch": 0.8796680497925311,
+      "grad_norm": 3.481823321693955,
+      "kl": 0.0869140625,
+      "learning_rate": 1.203319502074689e-07,
+      "loss": 0.0035,
+      "reward": 1.26953125,
+      "reward_std": 0.32253044843673706,
+      "rewards/accuracy_reward": 0.30078125,
+      "rewards/format_reward": 0.96875,
+      "step": 212
+    },
+    {
+      "completion_length": 95.96484375,
+      "epoch": 0.8838174273858921,
+      "grad_norm": 2.4104826815675646,
+      "kl": 0.0703125,
+      "learning_rate": 1.1618257261410788e-07,
+      "loss": 0.0028,
+      "reward": 1.1796875,
+      "reward_std": 0.251188188791275,
+      "rewards/accuracy_reward": 0.203125,
+      "rewards/format_reward": 0.9765625,
+      "step": 213
+    },
+    {
+      "completion_length": 98.62109375,
+      "epoch": 0.8879668049792531,
+      "grad_norm": 2.2792919284740303,
+      "kl": 0.07568359375,
+      "learning_rate": 1.1203319502074688e-07,
+      "loss": 0.003,
+      "reward": 1.3203125,
+      "reward_std": 0.31220242381095886,
+      "rewards/accuracy_reward": 0.3359375,
+      "rewards/format_reward": 0.984375,
+      "step": 214
+    },
+    {
+      "completion_length": 100.05078125,
+      "epoch": 0.8921161825726142,
+      "grad_norm": 4.276254804277667,
+      "kl": 0.08349609375,
+      "learning_rate": 1.0788381742738589e-07,
+      "loss": 0.0033,
+      "reward": 1.23828125,
+      "reward_std": 0.2676299810409546,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.97265625,
+      "step": 215
+    },
+    {
+      "completion_length": 93.05078125,
+      "epoch": 0.8962655601659751,
+      "grad_norm": 3.5229171032652435,
+      "kl": 0.0986328125,
+      "learning_rate": 1.037344398340249e-07,
+      "loss": 0.004,
+      "reward": 1.25,
+      "reward_std": 0.31738150119781494,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.9765625,
+      "step": 216
+    },
+    {
+      "completion_length": 94.16015625,
+      "epoch": 0.9004149377593361,
+      "grad_norm": 11.757057583856977,
+      "kl": 0.078125,
+      "learning_rate": 9.958506224066389e-08,
+      "loss": 0.0031,
+      "reward": 1.21875,
+      "reward_std": 0.24235805869102478,
+      "rewards/accuracy_reward": 0.23046875,
+      "rewards/format_reward": 0.98828125,
+      "step": 217
+    },
+    {
+      "completion_length": 85.44140625,
+      "epoch": 0.9045643153526971,
+      "grad_norm": 1.9091471037010437,
+      "kl": 0.103515625,
+      "learning_rate": 9.54356846473029e-08,
+      "loss": 0.0041,
+      "reward": 1.1953125,
+      "reward_std": 0.2225247025489807,
+      "rewards/accuracy_reward": 0.2109375,
+      "rewards/format_reward": 0.984375,
+      "step": 218
+    },
+    {
+      "completion_length": 96.31640625,
+      "epoch": 0.9087136929460581,
+      "grad_norm": 3.0098424472404544,
+      "kl": 0.07568359375,
+      "learning_rate": 9.128630705394189e-08,
+      "loss": 0.003,
+      "reward": 1.31640625,
+      "reward_std": 0.3151411712169647,
+      "rewards/accuracy_reward": 0.3359375,
+      "rewards/format_reward": 0.98046875,
+      "step": 219
+    },
+    {
+      "completion_length": 102.65234375,
+      "epoch": 0.9128630705394191,
+      "grad_norm": 4.643564268046483,
+      "kl": 0.0888671875,
+      "learning_rate": 8.71369294605809e-08,
+      "loss": 0.0036,
+      "reward": 1.15625,
+      "reward_std": 0.2785586714744568,
+      "rewards/accuracy_reward": 0.19921875,
+      "rewards/format_reward": 0.95703125,
+      "step": 220
+    },
+    {
+      "completion_length": 93.80859375,
+      "epoch": 0.91701244813278,
+      "grad_norm": 4.410133029194588,
+      "kl": 0.06982421875,
+      "learning_rate": 8.298755186721991e-08,
+      "loss": 0.0028,
+      "reward": 1.24609375,
+      "reward_std": 0.20357662439346313,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 1.0,
+      "step": 221
+    },
+    {
+      "completion_length": 101.15234375,
+      "epoch": 0.921161825726141,
+      "grad_norm": 2.904709180452436,
+      "kl": 0.08154296875,
+      "learning_rate": 7.883817427385891e-08,
+      "loss": 0.0033,
+      "reward": 1.19140625,
+      "reward_std": 0.2115919589996338,
+      "rewards/accuracy_reward": 0.21484375,
+      "rewards/format_reward": 0.9765625,
+      "step": 222
+    },
+    {
+      "completion_length": 96.94140625,
+      "epoch": 0.9253112033195021,
+      "grad_norm": 4.354536159655136,
+      "kl": 0.0751953125,
+      "learning_rate": 7.468879668049792e-08,
+      "loss": 0.003,
+      "reward": 1.24609375,
+      "reward_std": 0.21827393770217896,
+      "rewards/accuracy_reward": 0.26953125,
+      "rewards/format_reward": 0.9765625,
+      "step": 223
+    },
+    {
+      "completion_length": 93.47265625,
+      "epoch": 0.9294605809128631,
+      "grad_norm": 6.600006221648485,
+      "kl": 0.12451171875,
+      "learning_rate": 7.053941908713693e-08,
+      "loss": 0.005,
+      "reward": 1.15625,
+      "reward_std": 0.2866210341453552,
+      "rewards/accuracy_reward": 0.2109375,
+      "rewards/format_reward": 0.9453125,
+      "step": 224
+    },
+    {
+      "completion_length": 101.734375,
+      "epoch": 0.9336099585062241,
+      "grad_norm": 2.59933094148854,
+      "kl": 0.0810546875,
+      "learning_rate": 6.639004149377593e-08,
+      "loss": 0.0032,
+      "reward": 1.31640625,
+      "reward_std": 0.34125426411628723,
+      "rewards/accuracy_reward": 0.33984375,
+      "rewards/format_reward": 0.9765625,
+      "step": 225
+    },
+    {
+      "completion_length": 91.54296875,
+      "epoch": 0.9377593360995851,
+      "grad_norm": 2.8944936970951853,
+      "kl": 0.0947265625,
+      "learning_rate": 6.224066390041494e-08,
+      "loss": 0.0038,
+      "reward": 1.28515625,
+      "reward_std": 0.2791742980480194,
+      "rewards/accuracy_reward": 0.3046875,
+      "rewards/format_reward": 0.98046875,
+      "step": 226
+    },
+    {
+      "completion_length": 103.78125,
+      "epoch": 0.941908713692946,
+      "grad_norm": 4.590255163351441,
+      "kl": 0.0830078125,
+      "learning_rate": 5.809128630705394e-08,
+      "loss": 0.0033,
+      "reward": 1.25390625,
+      "reward_std": 0.3090318739414215,
+      "rewards/accuracy_reward": 0.29296875,
+      "rewards/format_reward": 0.9609375,
+      "step": 227
+    },
+    {
+      "completion_length": 94.26953125,
+      "epoch": 0.946058091286307,
+      "grad_norm": 2.1761507216115255,
+      "kl": 0.09423828125,
+      "learning_rate": 5.3941908713692946e-08,
+      "loss": 0.0038,
+      "reward": 1.2421875,
+      "reward_std": 0.23839645087718964,
+      "rewards/accuracy_reward": 0.28515625,
+      "rewards/format_reward": 0.95703125,
+      "step": 228
+    },
+    {
+      "completion_length": 97.3984375,
+      "epoch": 0.950207468879668,
+      "grad_norm": 4.778688881591497,
+      "kl": 0.09375,
+      "learning_rate": 4.979253112033194e-08,
+      "loss": 0.0037,
+      "reward": 1.2421875,
+      "reward_std": 0.24435341358184814,
+      "rewards/accuracy_reward": 0.265625,
+      "rewards/format_reward": 0.9765625,
+      "step": 229
+    },
+    {
+      "completion_length": 103.171875,
+      "epoch": 0.9543568464730291,
+      "grad_norm": 2.9856087811245935,
+      "kl": 0.08935546875,
+      "learning_rate": 4.5643153526970947e-08,
+      "loss": 0.0036,
+      "reward": 1.34765625,
+      "reward_std": 0.31364643573760986,
+      "rewards/accuracy_reward": 0.37890625,
+      "rewards/format_reward": 0.96875,
+      "step": 230
+    },
+    {
+      "completion_length": 96.609375,
+      "epoch": 0.9585062240663901,
+      "grad_norm": 4.58965397807011,
+      "kl": 0.076171875,
+      "learning_rate": 4.149377593360996e-08,
+      "loss": 0.0031,
+      "reward": 1.2265625,
+      "reward_std": 0.2816419303417206,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.98046875,
+      "step": 231
+    },
+    {
+      "completion_length": 106.4140625,
+      "epoch": 0.9626556016597511,
+      "grad_norm": 4.259706500288727,
+      "kl": 0.083984375,
+      "learning_rate": 3.734439834024896e-08,
+      "loss": 0.0034,
+      "reward": 1.19921875,
+      "reward_std": 0.2757681608200073,
+      "rewards/accuracy_reward": 0.24609375,
+      "rewards/format_reward": 0.953125,
+      "step": 232
+    },
+    {
+      "completion_length": 101.8203125,
+      "epoch": 0.966804979253112,
+      "grad_norm": 4.0534911383906085,
+      "kl": 0.078125,
+      "learning_rate": 3.3195020746887964e-08,
+      "loss": 0.0031,
+      "reward": 1.26953125,
+      "reward_std": 0.2769816219806671,
+      "rewards/accuracy_reward": 0.3125,
+      "rewards/format_reward": 0.95703125,
+      "step": 233
+    },
+    {
+      "completion_length": 102.25,
+      "epoch": 0.970954356846473,
+      "grad_norm": 4.9799320704360825,
+      "kl": 0.08837890625,
+      "learning_rate": 2.904564315352697e-08,
+      "loss": 0.0035,
+      "reward": 1.29296875,
+      "reward_std": 0.3803676962852478,
+      "rewards/accuracy_reward": 0.33984375,
+      "rewards/format_reward": 0.953125,
+      "step": 234
+    },
+    {
+      "completion_length": 84.4921875,
+      "epoch": 0.975103734439834,
+      "grad_norm": 7.36284248917279,
+      "kl": 0.11181640625,
+      "learning_rate": 2.489626556016597e-08,
+      "loss": 0.0045,
+      "reward": 1.23828125,
+      "reward_std": 0.26031655073165894,
+      "rewards/accuracy_reward": 0.26171875,
+      "rewards/format_reward": 0.9765625,
+      "step": 235
+    },
+    {
+      "completion_length": 101.4765625,
+      "epoch": 0.979253112033195,
+      "grad_norm": 4.233011578281721,
+      "kl": 0.080078125,
+      "learning_rate": 2.074688796680498e-08,
+      "loss": 0.0032,
+      "reward": 1.3203125,
+      "reward_std": 0.3218572735786438,
+      "rewards/accuracy_reward": 0.33984375,
+      "rewards/format_reward": 0.98046875,
+      "step": 236
+    },
+    {
+      "completion_length": 94.22265625,
+      "epoch": 0.983402489626556,
+      "grad_norm": 2.619642461579283,
+      "kl": 0.08837890625,
+      "learning_rate": 1.6597510373443982e-08,
+      "loss": 0.0035,
+      "reward": 1.2890625,
+      "reward_std": 0.32815423607826233,
+      "rewards/accuracy_reward": 0.3203125,
+      "rewards/format_reward": 0.96875,
+      "step": 237
+    },
+    {
+      "completion_length": 97.37890625,
+      "epoch": 0.9875518672199171,
+      "grad_norm": 3.912385420123859,
+      "kl": 0.095703125,
+      "learning_rate": 1.2448132780082986e-08,
+      "loss": 0.0038,
+      "reward": 1.2109375,
+      "reward_std": 0.2913069427013397,
+      "rewards/accuracy_reward": 0.2578125,
+      "rewards/format_reward": 0.953125,
+      "step": 238
+    },
+    {
+      "completion_length": 106.0234375,
+      "epoch": 0.991701244813278,
+      "grad_norm": 3.1256612206439294,
+      "kl": 0.08935546875,
+      "learning_rate": 8.298755186721991e-09,
+      "loss": 0.0036,
+      "reward": 1.2421875,
+      "reward_std": 0.2937677502632141,
+      "rewards/accuracy_reward": 0.2734375,
+      "rewards/format_reward": 0.96875,
+      "step": 239
+    },
+    {
+      "completion_length": 98.15625,
+      "epoch": 0.995850622406639,
+      "grad_norm": 3.5682820113500076,
+      "kl": 0.08740234375,
+      "learning_rate": 4.1493775933609955e-09,
+      "loss": 0.0035,
+      "reward": 1.29296875,
+      "reward_std": 0.3183509409427643,
+      "rewards/accuracy_reward": 0.3359375,
+      "rewards/format_reward": 0.95703125,
+      "step": 240
+    },
+    {
+      "completion_length": 77.55555725097656,
+      "epoch": 1.0,
+      "grad_norm": 4.43221819261977,
+      "kl": 0.12255859375,
+      "learning_rate": 0.0,
+      "loss": 0.004,
+      "reward": 1.3333333730697632,
+      "reward_std": 0.41147559881210327,
+      "rewards/accuracy_reward": 0.3333333432674408,
+      "rewards/format_reward": 1.0,
+      "step": 241
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 241,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-241/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29a51e69401bcc89579d8aeb02b6332a569eb27494642cc4cb4990bb76aa7073
+size 7672

checkpoint-241/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)

config.json

@@ -0,0 +1,48 @@
+{
+  "_name_or_path": "Qwen/Qwen2-VL-7B-Instruct",
+  "architectures": [
+    "Qwen2VLForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "image_token_id": 151655,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2_vl",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "mrope_section": [
+      16,
+      24,
+      24
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "in_chans": 3,
+    "model_type": "qwen2_vl",
+    "spatial_patch_size": 14,
+    "torch_dtype": "float32"
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "vision_token_id": 151654,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,15 @@
+{
+  "attn_implementation": "flash_attention_2",
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.01,
+  "top_k": 1,
+  "top_p": 0.001,
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false
+}