added 7B qwen model checkpoint

VideoSAVi-Qwen-7B-all-CLIP/added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

VideoSAVi-Qwen-7B-all-CLIP/config.json

@@ -0,0 +1,78 @@
+{
+  "_name_or_path": "/scratch/ykulka10/llava-next-interleave-qwen-7b",
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "freeze_mm_mlp_adapter": false,
+  "freeze_mm_vision_resampler": false,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "image_aspect_ratio": "anyres",
+  "image_crop_resolution": 384,
+  "image_grid_pinpoints": [
+    [
+      384,
+      768
+    ],
+    [
+      768,
+      384
+    ],
+    [
+      768,
+      768
+    ],
+    [
+      1152,
+      384
+    ],
+    [
+      384,
+      1152
+    ]
+  ],
+  "image_split_resolution": 384,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "mm_hidden_size": 1152,
+  "mm_patch_merge_type": "spatial_unpad",
+  "mm_projector_lr": null,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": "spatial_pool",
+  "mm_spatial_pool_mode": "average",
+  "mm_spatial_pool_out_channels": 1024,
+  "mm_spatial_pool_stride": 2,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "google/siglip-so400m-patch14-384",
+  "mm_vision_tower_lr": null,
+  "model_type": "llava_qwen",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.40.0.dev0",
+  "tune_mm_mlp_adapter": false,
+  "tune_mm_vision_resampler": false,
+  "unfreeze_mm_vision_tower": false,
+  "use_cache": false,
+  "use_mm_proj": true,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": null,
+  "vocab_size": 151936
+}

VideoSAVi-Qwen-7B-all-CLIP/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,
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.40.0.dev0"
+}

VideoSAVi-Qwen-7B-all-CLIP/latest

@@ -0,0 +1 @@
+global_step150

VideoSAVi-Qwen-7B-all-CLIP/model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df108c72a308484bcdfca64aa296b82235c3fc2c6c3143f0eeae2de41f3c6edd
+size 4988490704

VideoSAVi-Qwen-7B-all-CLIP/model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5bdb72ab18ee283f30d45a4e560ac6d5b93bacc39301c09201602e0a1791097
+size 4981244328

VideoSAVi-Qwen-7B-all-CLIP/model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:835847a1b03e677f4029edc07495aa1e4f09d1263aa7a51cf0934cf920e3b2dc
+size 4998695864

VideoSAVi-Qwen-7B-all-CLIP/model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af19d973c7c37face9edeeffda8bbd9a0c92a26820beb67f52f7e4957ad7e756
+size 1312841000

VideoSAVi-Qwen-7B-all-CLIP/model.safetensors.index.json

@@ -0,0 +1,820 @@
+{
+  "metadata": {
+    "total_size": 16281160768
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.image_newline": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-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-00002-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-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-00002-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-00002-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-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-00001-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.mm_projector.0.bias": "model-00004-of-00004.safetensors",
+    "model.mm_projector.0.weight": "model-00004-of-00004.safetensors",
+    "model.mm_projector.2.bias": "model-00004-of-00004.safetensors",
+    "model.mm_projector.2.weight": "model-00004-of-00004.safetensors",
+    "model.norm.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.embeddings.patch_embedding.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.embeddings.patch_embedding.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.embeddings.position_embedding.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.layer_norm1.bias": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.layer_norm1.weight": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.layer_norm2.bias": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.layer_norm2.weight": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.mlp.fc1.bias": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.mlp.fc1.weight": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.mlp.fc2.bias": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.mlp.fc2.weight": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.layer_norm1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.layer_norm1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.layer_norm2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.layer_norm2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.mlp.fc1.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.mlp.fc1.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.mlp.fc2.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.mlp.fc2.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.post_layernorm.bias": "model-00004-of-00004.safetensors",
+    "model.vision_tower.vision_tower.vision_model.post_layernorm.weight": "model-00004-of-00004.safetensors"
+  }
+}

VideoSAVi-Qwen-7B-all-CLIP/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d63731b330e0f735fe2992a8913e78081d07ee0cc42996b9adee30db79fee95d
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:015b5f15f8fe064843b4701455431376f593b118ea16b386ec0ec0142e2cbbcf
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07bbb9859c307f6c529da2ea8c74f6539fc49961386200cf5a70e0dc570fef06
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:723c45e188af3ade31c8856c59064462207d117a2305dc3205aaaf999d43a9cd
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b734507b167ad50518ddc4e1922f10d7e14835aaaf6667514967c9b7f4712784
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3c678962f778817757fece84872578ae799549ead8da5bc26db230adae0409a
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cada924bbf45d9df603cc25552969eaec3f62aa68d9876a08b71f24621fffeae
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aca0d2229c40d555559f523f0824608f2090c50c4fa537408cc110079362a469
+size 15984

VideoSAVi-Qwen-7B-all-CLIP/special_tokens_map.json

@@ -0,0 +1,20 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "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
+  }
+}

VideoSAVi-Qwen-7B-all-CLIP/tokenizer_config.json

@@ -0,0 +1,44 @@
+{
+  "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
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 2048,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

VideoSAVi-Qwen-7B-all-CLIP/trainer_state.json

@@ -0,0 +1,2721 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.0340522133938706,
+  "eval_steps": 500,
+  "global_step": 150,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "grad_norm": 53.13322448730469,
+      "learning_rate": 3.7821482602118e-10,
+      "logps/chosen": -45.965152740478516,
+      "logps/rejected": -43.26817321777344,
+      "loss": 0.6931,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.4698270559310913,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -45.965152740478516,
+      "ref_logps/rejected": -43.26817321777344,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 42.47979736328125,
+      "learning_rate": 7.5642965204236e-10,
+      "logps/chosen": -40.02202606201172,
+      "logps/rejected": -32.612422943115234,
+      "loss": 0.6931,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.39600855112075806,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -40.02202606201172,
+      "ref_logps/rejected": -32.612422943115234,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 41.06604766845703,
+      "learning_rate": 1.13464447806354e-09,
+      "logps/chosen": -40.55537796020508,
+      "logps/rejected": -34.559226989746094,
+      "loss": 0.6953,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.42144468426704407,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -40.522464752197266,
+      "ref_logps/rejected": -34.568199157714844,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": -0.0032914162147790194,
+      "rewards/margins": -0.0041886805556714535,
+      "rewards/rejected": 0.0008972645155154169,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 37.634681701660156,
+      "learning_rate": 1.51285930408472e-09,
+      "logps/chosen": -30.457353591918945,
+      "logps/rejected": -35.24971389770508,
+      "loss": 0.698,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.7412044405937195,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -30.345169067382812,
+      "ref_logps/rejected": -35.23454284667969,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": -0.011218691244721413,
+      "rewards/margins": -0.009701800532639027,
+      "rewards/rejected": -0.0015168904792517424,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 42.97514724731445,
+      "learning_rate": 1.8910741301059e-09,
+      "logps/chosen": -37.70698547363281,
+      "logps/rejected": -23.77716064453125,
+      "loss": 0.6938,
+      "losses/dpo": 0.7050856947898865,
+      "losses/sft": 0.4762720763683319,
+      "losses/total": 0.7050856947898865,
+      "ref_logps/chosen": -37.755462646484375,
+      "ref_logps/rejected": -23.837926864624023,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.0048479679971933365,
+      "rewards/margins": -0.0012286424171179533,
+      "rewards/rejected": 0.0060766106471419334,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 30.40886116027832,
+      "learning_rate": 2.26928895612708e-09,
+      "logps/chosen": -19.241403579711914,
+      "logps/rejected": -21.918075561523438,
+      "loss": 0.6937,
+      "losses/dpo": 0.6973899602890015,
+      "losses/sft": 0.09948720037937164,
+      "losses/total": 0.6973899602890015,
+      "ref_logps/chosen": -19.206815719604492,
+      "ref_logps/rejected": -21.89419174194336,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.0034588337875902653,
+      "rewards/margins": -0.00107060675509274,
+      "rewards/rejected": -0.0023882267996668816,
+      "step": 6
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 43.89079284667969,
+      "learning_rate": 2.64750378214826e-09,
+      "logps/chosen": -35.801856994628906,
+      "logps/rejected": -36.446815490722656,
+      "loss": 0.6898,
+      "losses/dpo": 0.6871867775917053,
+      "losses/sft": 1.0996204614639282,
+      "losses/total": 0.6871867775917053,
+      "ref_logps/chosen": -35.87968063354492,
+      "ref_logps/rejected": -36.45803451538086,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.007782352156937122,
+      "rewards/margins": 0.006660759914666414,
+      "rewards/rejected": 0.0011215927079319954,
+      "step": 7
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 46.40654754638672,
+      "learning_rate": 3.02571860816944e-09,
+      "logps/chosen": -23.76507568359375,
+      "logps/rejected": -49.3902473449707,
+      "loss": 0.6964,
+      "losses/dpo": 0.6929582357406616,
+      "losses/sft": 0.39775577187538147,
+      "losses/total": 0.6929582357406616,
+      "ref_logps/chosen": -23.763172149658203,
+      "ref_logps/rejected": -49.45380401611328,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.000190228340215981,
+      "rewards/margins": -0.006545871961861849,
+      "rewards/rejected": 0.006355643272399902,
+      "step": 8
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 41.16609191894531,
+      "learning_rate": 3.40393343419062e-09,
+      "logps/chosen": -46.34278869628906,
+      "logps/rejected": -39.25700378417969,
+      "loss": 0.6854,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.9487500190734863,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -46.34722137451172,
+      "ref_logps/rejected": -39.10374450683594,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.00044331513345241547,
+      "rewards/margins": 0.015769505873322487,
+      "rewards/rejected": -0.015326190739870071,
+      "step": 9
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 36.190269470214844,
+      "learning_rate": 3.7821482602118e-09,
+      "logps/chosen": -25.09479331970215,
+      "logps/rejected": -39.543060302734375,
+      "loss": 0.6939,
+      "losses/dpo": 0.6988048553466797,
+      "losses/sft": 1.1739815473556519,
+      "losses/total": 0.6988048553466797,
+      "ref_logps/chosen": -24.979263305664062,
+      "ref_logps/rejected": -39.441261291503906,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.011553037911653519,
+      "rewards/margins": -0.0013730167411267757,
+      "rewards/rejected": -0.010180020704865456,
+      "step": 10
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 42.968685150146484,
+      "learning_rate": 4.16036308623298e-09,
+      "logps/chosen": -45.688053131103516,
+      "logps/rejected": -31.156063079833984,
+      "loss": 0.689,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 1.371110439300537,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -45.754638671875,
+      "ref_logps/rejected": -31.138751983642578,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0066588278859853745,
+      "rewards/margins": 0.008389795199036598,
+      "rewards/rejected": -0.0017309663817286491,
+      "step": 11
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 34.414268493652344,
+      "learning_rate": 4.53857791225416e-09,
+      "logps/chosen": -36.18522644042969,
+      "logps/rejected": -29.030437469482422,
+      "loss": 0.6954,
+      "losses/dpo": 0.6858934760093689,
+      "losses/sft": 0.42428943514823914,
+      "losses/total": 0.6858934760093689,
+      "ref_logps/chosen": -36.183631896972656,
+      "ref_logps/rejected": -29.07280731201172,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.00015949038788676262,
+      "rewards/margins": -0.004396378993988037,
+      "rewards/rejected": 0.004236889071762562,
+      "step": 12
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 42.252872467041016,
+      "learning_rate": 4.9167927382753406e-09,
+      "logps/chosen": -30.722270965576172,
+      "logps/rejected": -52.374778747558594,
+      "loss": 0.6921,
+      "losses/dpo": 0.6986141800880432,
+      "losses/sft": 0.7042618989944458,
+      "losses/total": 0.6986141800880432,
+      "ref_logps/chosen": -30.702848434448242,
+      "ref_logps/rejected": -52.333797454833984,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.0019423721823841333,
+      "rewards/margins": 0.0021560671739280224,
+      "rewards/rejected": -0.004098439123481512,
+      "step": 13
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 52.068031311035156,
+      "learning_rate": 5.29500756429652e-09,
+      "logps/chosen": -58.0643310546875,
+      "logps/rejected": -40.236019134521484,
+      "loss": 0.6829,
+      "losses/dpo": 0.7056142091751099,
+      "losses/sft": 0.8749198317527771,
+      "losses/total": 0.7056142091751099,
+      "ref_logps/chosen": -58.133060455322266,
+      "ref_logps/rejected": -40.09267807006836,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.00687251053750515,
+      "rewards/margins": 0.021206431090831757,
+      "rewards/rejected": -0.014333921484649181,
+      "step": 14
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 33.8675651550293,
+      "learning_rate": 5.6732223903177e-09,
+      "logps/chosen": -28.738069534301758,
+      "logps/rejected": -32.7332649230957,
+      "loss": 0.6882,
+      "losses/dpo": 0.6855644583702087,
+      "losses/sft": 0.6099794507026672,
+      "losses/total": 0.6855644583702087,
+      "ref_logps/chosen": -28.814668655395508,
+      "ref_logps/rejected": -32.710052490234375,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.007659816648811102,
+      "rewards/margins": 0.009981178678572178,
+      "rewards/rejected": -0.0023213629610836506,
+      "step": 15
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 50.59415054321289,
+      "learning_rate": 6.05143721633888e-09,
+      "logps/chosen": -63.415924072265625,
+      "logps/rejected": -34.18522644042969,
+      "loss": 0.687,
+      "losses/dpo": 0.6864607930183411,
+      "losses/sft": 0.6457326412200928,
+      "losses/total": 0.6864607930183411,
+      "ref_logps/chosen": -63.533287048339844,
+      "ref_logps/rejected": -34.1763801574707,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.011736572720110416,
+      "rewards/margins": 0.01262127235531807,
+      "rewards/rejected": -0.0008846997516229749,
+      "step": 16
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 31.88450050354004,
+      "learning_rate": 6.429652042360061e-09,
+      "logps/chosen": -19.916322708129883,
+      "logps/rejected": -28.502450942993164,
+      "loss": 0.7005,
+      "losses/dpo": 0.6912732720375061,
+      "losses/sft": 0.2310972660779953,
+      "losses/total": 0.6912732720375061,
+      "ref_logps/chosen": -19.821392059326172,
+      "ref_logps/rejected": -28.552419662475586,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.00949319638311863,
+      "rewards/margins": -0.014490035362541676,
+      "rewards/rejected": 0.004996838979423046,
+      "step": 17
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 44.439125061035156,
+      "learning_rate": 6.80786686838124e-09,
+      "logps/chosen": -21.987422943115234,
+      "logps/rejected": -52.99866485595703,
+      "loss": 0.6919,
+      "losses/dpo": 0.6910991072654724,
+      "losses/sft": 0.6097382307052612,
+      "losses/total": 0.6910991072654724,
+      "ref_logps/chosen": -22.110370635986328,
+      "ref_logps/rejected": -53.094139099121094,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.012294775806367397,
+      "rewards/margins": 0.0027476726099848747,
+      "rewards/rejected": 0.009547103196382523,
+      "step": 18
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 47.79739761352539,
+      "learning_rate": 7.18608169440242e-09,
+      "logps/chosen": -58.563255310058594,
+      "logps/rejected": -41.6775016784668,
+      "loss": 0.6919,
+      "losses/dpo": 0.692899763584137,
+      "losses/sft": 0.41986706852912903,
+      "losses/total": 0.692899763584137,
+      "ref_logps/chosen": -58.672828674316406,
+      "ref_logps/rejected": -41.761592864990234,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.010957282967865467,
+      "rewards/margins": 0.0025482713244855404,
+      "rewards/rejected": 0.00840901117771864,
+      "step": 19
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 48.81391906738281,
+      "learning_rate": 7.5642965204236e-09,
+      "logps/chosen": -43.01997375488281,
+      "logps/rejected": -42.654544830322266,
+      "loss": 0.6909,
+      "losses/dpo": 0.7070827484130859,
+      "losses/sft": 0.4298737943172455,
+      "losses/total": 0.7070827484130859,
+      "ref_logps/chosen": -43.05825424194336,
+      "ref_logps/rejected": -42.64590072631836,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.0038282987661659718,
+      "rewards/margins": 0.004692500922828913,
+      "rewards/rejected": -0.0008642016910016537,
+      "step": 20
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 45.92574691772461,
+      "learning_rate": 7.94251134644478e-09,
+      "logps/chosen": -47.82428741455078,
+      "logps/rejected": -28.918079376220703,
+      "loss": 0.6931,
+      "losses/dpo": 0.6990569829940796,
+      "losses/sft": 0.884921669960022,
+      "losses/total": 0.6990569829940796,
+      "ref_logps/chosen": -47.84720230102539,
+      "ref_logps/rejected": -28.938232421875,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.0022917508613318205,
+      "rewards/margins": 0.00027650000993162394,
+      "rewards/rejected": 0.0020152509678155184,
+      "step": 21
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 33.18178176879883,
+      "learning_rate": 8.32072617246596e-09,
+      "logps/chosen": -29.66305160522461,
+      "logps/rejected": -29.4906063079834,
+      "loss": 0.6904,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.32695040106773376,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -29.655746459960938,
+      "ref_logps/rejected": -29.426069259643555,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0007306460756808519,
+      "rewards/margins": 0.005722904112190008,
+      "rewards/rejected": -0.0064535499550402164,
+      "step": 22
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 44.96432113647461,
+      "learning_rate": 8.69894099848714e-09,
+      "logps/chosen": -44.383628845214844,
+      "logps/rejected": -30.62492561340332,
+      "loss": 0.696,
+      "losses/dpo": 0.7036435008049011,
+      "losses/sft": 0.9137359261512756,
+      "losses/total": 0.7036435008049011,
+      "ref_logps/chosen": -44.326786041259766,
+      "ref_logps/rejected": -30.624820709228516,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0056844474747776985,
+      "rewards/margins": -0.0056739808060228825,
+      "rewards/rejected": -1.0466552339494228e-05,
+      "step": 23
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 42.07050323486328,
+      "learning_rate": 9.07715582450832e-09,
+      "logps/chosen": -42.034481048583984,
+      "logps/rejected": -36.58326721191406,
+      "loss": 0.7022,
+      "losses/dpo": 0.7029128074645996,
+      "losses/sft": 1.0267704725265503,
+      "losses/total": 0.7029128074645996,
+      "ref_logps/chosen": -41.85973358154297,
+      "ref_logps/rejected": -36.588218688964844,
+      "rewards/accuracies": 0.125,
+      "rewards/chosen": -0.017474675551056862,
+      "rewards/margins": -0.017969846725463867,
+      "rewards/rejected": 0.0004951715236529708,
+      "step": 24
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 40.44936752319336,
+      "learning_rate": 9.4553706505295e-09,
+      "logps/chosen": -28.277244567871094,
+      "logps/rejected": -38.51260757446289,
+      "loss": 0.6901,
+      "losses/dpo": 0.6949249505996704,
+      "losses/sft": 0.8017578721046448,
+      "losses/total": 0.6949249505996704,
+      "ref_logps/chosen": -28.28316879272461,
+      "ref_logps/rejected": -38.45671081542969,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.0005923390854150057,
+      "rewards/margins": 0.0061820391565561295,
+      "rewards/rejected": -0.005589699372649193,
+      "step": 25
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 39.54838562011719,
+      "learning_rate": 9.833585476550681e-09,
+      "logps/chosen": -39.214866638183594,
+      "logps/rejected": -35.28057098388672,
+      "loss": 0.6906,
+      "losses/dpo": 0.6930829286575317,
+      "losses/sft": 0.2879458963871002,
+      "losses/total": 0.6930829286575317,
+      "ref_logps/chosen": -39.14422607421875,
+      "ref_logps/rejected": -35.15850830078125,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.007063663098961115,
+      "rewards/margins": 0.005142513196915388,
+      "rewards/rejected": -0.012206176295876503,
+      "step": 26
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 46.594966888427734,
+      "learning_rate": 1.021180030257186e-08,
+      "logps/chosen": -52.03388595581055,
+      "logps/rejected": -38.81571578979492,
+      "loss": 0.7028,
+      "losses/dpo": 0.7005717158317566,
+      "losses/sft": 0.19097577035427094,
+      "losses/total": 0.7005717158317566,
+      "ref_logps/chosen": -51.97730255126953,
+      "ref_logps/rejected": -38.94976043701172,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.005658918991684914,
+      "rewards/margins": -0.01906323991715908,
+      "rewards/rejected": 0.013404322788119316,
+      "step": 27
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 48.950103759765625,
+      "learning_rate": 1.059001512859304e-08,
+      "logps/chosen": -46.73711395263672,
+      "logps/rejected": -59.71778869628906,
+      "loss": 0.6948,
+      "losses/dpo": 0.7069879174232483,
+      "losses/sft": 1.273395299911499,
+      "losses/total": 0.7069879174232483,
+      "ref_logps/chosen": -46.688907623291016,
+      "ref_logps/rejected": -59.699073791503906,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.004820322617888451,
+      "rewards/margins": -0.002948999870568514,
+      "rewards/rejected": -0.0018713227473199368,
+      "step": 28
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 44.046485900878906,
+      "learning_rate": 1.096822995461422e-08,
+      "logps/chosen": -43.880401611328125,
+      "logps/rejected": -40.68788146972656,
+      "loss": 0.6967,
+      "losses/dpo": 0.7051054239273071,
+      "losses/sft": 0.8562471866607666,
+      "losses/total": 0.7051054239273071,
+      "ref_logps/chosen": -43.754669189453125,
+      "ref_logps/rejected": -40.6314811706543,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.012573599815368652,
+      "rewards/margins": -0.006933570373803377,
+      "rewards/rejected": -0.0056400299072265625,
+      "step": 29
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 22.69830894470215,
+      "learning_rate": 1.13464447806354e-08,
+      "logps/chosen": -19.23567008972168,
+      "logps/rejected": -17.581375122070312,
+      "loss": 0.6965,
+      "losses/dpo": 0.7024092674255371,
+      "losses/sft": 0.8522212505340576,
+      "losses/total": 0.7024092674255371,
+      "ref_logps/chosen": -19.222492218017578,
+      "ref_logps/rejected": -17.634857177734375,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.0013178824447095394,
+      "rewards/margins": -0.006665962748229504,
+      "rewards/rejected": 0.0053480807691812515,
+      "step": 30
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 48.038639068603516,
+      "learning_rate": 1.172465960665658e-08,
+      "logps/chosen": -57.04039001464844,
+      "logps/rejected": -38.704063415527344,
+      "loss": 0.6936,
+      "losses/dpo": 0.6908040046691895,
+      "losses/sft": 0.8999913930892944,
+      "losses/total": 0.6908040046691895,
+      "ref_logps/chosen": -57.11170196533203,
+      "ref_logps/rejected": -38.78229522705078,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.007131719961762428,
+      "rewards/margins": -0.0006912816315889359,
+      "rewards/rejected": 0.007823002524673939,
+      "step": 31
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 49.33280944824219,
+      "learning_rate": 1.210287443267776e-08,
+      "logps/chosen": -52.908721923828125,
+      "logps/rejected": -51.61956024169922,
+      "loss": 0.6959,
+      "losses/dpo": 0.6972215175628662,
+      "losses/sft": 0.24405121803283691,
+      "losses/total": 0.6972215175628662,
+      "ref_logps/chosen": -52.914344787597656,
+      "ref_logps/rejected": -51.67469787597656,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.000563180074095726,
+      "rewards/margins": -0.004950487520545721,
+      "rewards/rejected": 0.005513668991625309,
+      "step": 32
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 42.91134262084961,
+      "learning_rate": 1.2481089258698941e-08,
+      "logps/chosen": -29.194244384765625,
+      "logps/rejected": -46.12828826904297,
+      "loss": 0.6912,
+      "losses/dpo": 0.6876184344291687,
+      "losses/sft": 0.43061476945877075,
+      "losses/total": 0.6876184344291687,
+      "ref_logps/chosen": -29.302093505859375,
+      "ref_logps/rejected": -46.194671630859375,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.010784798301756382,
+      "rewards/margins": 0.004146045073866844,
+      "rewards/rejected": 0.006638753227889538,
+      "step": 33
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 38.4871826171875,
+      "learning_rate": 1.2859304084720122e-08,
+      "logps/chosen": -29.756376266479492,
+      "logps/rejected": -22.08219337463379,
+      "loss": 0.6916,
+      "losses/dpo": 0.6908648014068604,
+      "losses/sft": 0.6852660179138184,
+      "losses/total": 0.6908648014068604,
+      "ref_logps/chosen": -29.772045135498047,
+      "ref_logps/rejected": -22.064563751220703,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.0015667201951146126,
+      "rewards/margins": 0.0033298381604254246,
+      "rewards/rejected": -0.0017631174996495247,
+      "step": 34
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 46.963951110839844,
+      "learning_rate": 1.3237518910741299e-08,
+      "logps/chosen": -35.22114181518555,
+      "logps/rejected": -48.63469314575195,
+      "loss": 0.7005,
+      "losses/dpo": 0.6885900497436523,
+      "losses/sft": 0.353080153465271,
+      "losses/total": 0.6885900497436523,
+      "ref_logps/chosen": -35.19417190551758,
+      "ref_logps/rejected": -48.75211715698242,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.002697205636650324,
+      "rewards/margins": -0.014439558610320091,
+      "rewards/rejected": 0.011742353439331055,
+      "step": 35
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 36.61310577392578,
+      "learning_rate": 1.361573373676248e-08,
+      "logps/chosen": -27.793731689453125,
+      "logps/rejected": -31.242351531982422,
+      "loss": 0.6951,
+      "losses/dpo": 0.7104878425598145,
+      "losses/sft": 0.9098966717720032,
+      "losses/total": 0.7104878425598145,
+      "ref_logps/chosen": -27.74094581604004,
+      "ref_logps/rejected": -31.22858428955078,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.005278551485389471,
+      "rewards/margins": -0.0039020059630274773,
+      "rewards/rejected": -0.0013765455223619938,
+      "step": 36
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 34.854759216308594,
+      "learning_rate": 1.399394856278366e-08,
+      "logps/chosen": -27.95212745666504,
+      "logps/rejected": -27.223896026611328,
+      "loss": 0.6904,
+      "losses/dpo": 0.6964800357818604,
+      "losses/sft": 0.20950362086296082,
+      "losses/total": 0.6964800357818604,
+      "ref_logps/chosen": -27.97368621826172,
+      "ref_logps/rejected": -27.19019889831543,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.002155975205823779,
+      "rewards/margins": 0.00552578829228878,
+      "rewards/rejected": -0.0033698140177875757,
+      "step": 37
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 40.405250549316406,
+      "learning_rate": 1.437216338880484e-08,
+      "logps/chosen": -37.7248649597168,
+      "logps/rejected": -44.841819763183594,
+      "loss": 0.6867,
+      "losses/dpo": 0.681260347366333,
+      "losses/sft": 1.189764380455017,
+      "losses/total": 0.681260347366333,
+      "ref_logps/chosen": -37.73345184326172,
+      "ref_logps/rejected": -44.71831130981445,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.0008587008342146873,
+      "rewards/margins": 0.013209670782089233,
+      "rewards/rejected": -0.01235097087919712,
+      "step": 38
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 49.105464935302734,
+      "learning_rate": 1.4750378214826021e-08,
+      "logps/chosen": -46.39025115966797,
+      "logps/rejected": -49.15430450439453,
+      "loss": 0.6952,
+      "losses/dpo": 0.6926788687705994,
+      "losses/sft": 1.5534383058547974,
+      "losses/total": 0.6926788687705994,
+      "ref_logps/chosen": -46.47325897216797,
+      "ref_logps/rejected": -49.276241302490234,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.008300865069031715,
+      "rewards/margins": -0.0038922191597521305,
+      "rewards/rejected": 0.012193083763122559,
+      "step": 39
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 28.91716194152832,
+      "learning_rate": 1.51285930408472e-08,
+      "logps/chosen": -29.146198272705078,
+      "logps/rejected": -20.899898529052734,
+      "loss": 0.6876,
+      "losses/dpo": 0.6868137717247009,
+      "losses/sft": 0.9044435620307922,
+      "losses/total": 0.6868137717247009,
+      "ref_logps/chosen": -29.24700164794922,
+      "ref_logps/rejected": -20.889738082885742,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.010080337524414062,
+      "rewards/margins": 0.011096293106675148,
+      "rewards/rejected": -0.0010159553494304419,
+      "step": 40
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 50.75545120239258,
+      "learning_rate": 1.550680786686838e-08,
+      "logps/chosen": -54.49714660644531,
+      "logps/rejected": -42.79107666015625,
+      "loss": 0.6764,
+      "losses/dpo": 0.6895805597305298,
+      "losses/sft": 0.8109897971153259,
+      "losses/total": 0.6895805597305298,
+      "ref_logps/chosen": -54.59727478027344,
+      "ref_logps/rejected": -42.551483154296875,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.010012972168624401,
+      "rewards/margins": 0.03397263213992119,
+      "rewards/rejected": -0.023959660902619362,
+      "step": 41
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 47.54016876220703,
+      "learning_rate": 1.588502269288956e-08,
+      "logps/chosen": -46.886924743652344,
+      "logps/rejected": -29.53119659423828,
+      "loss": 0.6922,
+      "losses/dpo": 0.6936006546020508,
+      "losses/sft": 0.5564570426940918,
+      "losses/total": 0.6936006546020508,
+      "ref_logps/chosen": -46.89942169189453,
+      "ref_logps/rejected": -29.524349212646484,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.0012490153312683105,
+      "rewards/margins": 0.0019336579134687781,
+      "rewards/rejected": -0.0006846425822004676,
+      "step": 42
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 50.0217170715332,
+      "learning_rate": 1.626323751891074e-08,
+      "logps/chosen": -29.848779678344727,
+      "logps/rejected": -59.39533615112305,
+      "loss": 0.7001,
+      "losses/dpo": 0.7061429619789124,
+      "losses/sft": 1.1746078729629517,
+      "losses/total": 0.7061429619789124,
+      "ref_logps/chosen": -29.843900680541992,
+      "ref_logps/rejected": -59.52110290527344,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.00048793526366353035,
+      "rewards/margins": -0.01306464709341526,
+      "rewards/rejected": 0.012576711364090443,
+      "step": 43
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 33.6051139831543,
+      "learning_rate": 1.664145234493192e-08,
+      "logps/chosen": -33.32111740112305,
+      "logps/rejected": -20.631572723388672,
+      "loss": 0.6905,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 1.1099908351898193,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -33.38194274902344,
+      "ref_logps/rejected": -20.636655807495117,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.006082272157073021,
+      "rewards/margins": 0.0055738575756549835,
+      "rewards/rejected": 0.0005084158619865775,
+      "step": 44
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 44.24886703491211,
+      "learning_rate": 1.70196671709531e-08,
+      "logps/chosen": -72.52376556396484,
+      "logps/rejected": -68.03264617919922,
+      "loss": 0.6977,
+      "losses/dpo": 0.7006234526634216,
+      "losses/sft": 0.5646258592605591,
+      "losses/total": 0.7006234526634216,
+      "ref_logps/chosen": -72.54812622070312,
+      "ref_logps/rejected": -68.1447982788086,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.002435469999909401,
+      "rewards/margins": -0.008779918774962425,
+      "rewards/rejected": 0.011215388774871826,
+      "step": 45
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 28.697980880737305,
+      "learning_rate": 1.739788199697428e-08,
+      "logps/chosen": -23.719348907470703,
+      "logps/rejected": -22.18777847290039,
+      "loss": 0.6999,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.4597294330596924,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -23.706619262695312,
+      "ref_logps/rejected": -22.308765411376953,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": -0.0012728453148156404,
+      "rewards/margins": -0.013371658511459827,
+      "rewards/rejected": 0.012098812498152256,
+      "step": 46
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 39.397056579589844,
+      "learning_rate": 1.777609682299546e-08,
+      "logps/chosen": -30.32215118408203,
+      "logps/rejected": -36.397674560546875,
+      "loss": 0.6964,
+      "losses/dpo": 0.6941532492637634,
+      "losses/sft": 0.9367761015892029,
+      "losses/total": 0.6941532492637634,
+      "ref_logps/chosen": -30.29210090637207,
+      "ref_logps/rejected": -36.431209564208984,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0030048727057874203,
+      "rewards/margins": -0.006358248647302389,
+      "rewards/rejected": 0.0033533747773617506,
+      "step": 47
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 39.70116424560547,
+      "learning_rate": 1.815431164901664e-08,
+      "logps/chosen": -36.93006896972656,
+      "logps/rejected": -40.943641662597656,
+      "loss": 0.6973,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 1.1891580820083618,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -36.934635162353516,
+      "ref_logps/rejected": -41.03041458129883,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.00045659602619707584,
+      "rewards/margins": -0.00822056457400322,
+      "rewards/rejected": 0.008677160367369652,
+      "step": 48
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 30.530561447143555,
+      "learning_rate": 1.853252647503782e-08,
+      "logps/chosen": -23.064416885375977,
+      "logps/rejected": -26.3908634185791,
+      "loss": 0.6978,
+      "losses/dpo": 0.6976159811019897,
+      "losses/sft": 0.7396925091743469,
+      "losses/total": 0.6976159811019897,
+      "ref_logps/chosen": -23.026689529418945,
+      "ref_logps/rejected": -26.44501495361328,
+      "rewards/accuracies": 0.125,
+      "rewards/chosen": -0.0037726524751633406,
+      "rewards/margins": -0.009187883697450161,
+      "rewards/rejected": 0.005415231455117464,
+      "step": 49
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 43.97072982788086,
+      "learning_rate": 1.8910741301059e-08,
+      "logps/chosen": -41.17816162109375,
+      "logps/rejected": -53.586917877197266,
+      "loss": 0.6952,
+      "losses/dpo": 0.7102349400520325,
+      "losses/sft": 0.7952568531036377,
+      "losses/total": 0.7102349400520325,
+      "ref_logps/chosen": -41.27253723144531,
+      "ref_logps/rejected": -53.72220230102539,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.00943751260638237,
+      "rewards/margins": -0.004090785980224609,
+      "rewards/rejected": 0.013528299517929554,
+      "step": 50
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 45.2219123840332,
+      "learning_rate": 1.9288956127080182e-08,
+      "logps/chosen": -30.593463897705078,
+      "logps/rejected": -29.29912567138672,
+      "loss": 0.7039,
+      "losses/dpo": 0.6956669688224792,
+      "losses/sft": 0.34454602003097534,
+      "losses/total": 0.6956669688224792,
+      "ref_logps/chosen": -30.442913055419922,
+      "ref_logps/rejected": -29.360437393188477,
+      "rewards/accuracies": 0.125,
+      "rewards/chosen": -0.015054965391755104,
+      "rewards/margins": -0.021185901015996933,
+      "rewards/rejected": 0.00613093376159668,
+      "step": 51
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 43.098976135253906,
+      "learning_rate": 1.9667170953101362e-08,
+      "logps/chosen": -45.34044647216797,
+      "logps/rejected": -39.2700309753418,
+      "loss": 0.6935,
+      "losses/dpo": 0.688653290271759,
+      "losses/sft": 0.6188009977340698,
+      "losses/total": 0.688653290271759,
+      "ref_logps/chosen": -45.37505340576172,
+      "ref_logps/rejected": -39.31009292602539,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.0034605623222887516,
+      "rewards/margins": -0.0005456925136968493,
+      "rewards/rejected": 0.004006254021078348,
+      "step": 52
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 40.16943359375,
+      "learning_rate": 2.0045385779122543e-08,
+      "logps/chosen": -31.46408462524414,
+      "logps/rejected": -26.625892639160156,
+      "loss": 0.6952,
+      "losses/dpo": 0.6910169720649719,
+      "losses/sft": 0.48614105582237244,
+      "losses/total": 0.6910169720649719,
+      "ref_logps/chosen": -31.48444938659668,
+      "ref_logps/rejected": -26.686222076416016,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.002036535646766424,
+      "rewards/margins": -0.003996384330093861,
+      "rewards/rejected": 0.006032920442521572,
+      "step": 53
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 49.609840393066406,
+      "learning_rate": 2.042360060514372e-08,
+      "logps/chosen": -43.83415985107422,
+      "logps/rejected": -45.29457092285156,
+      "loss": 0.6925,
+      "losses/dpo": 0.702146053314209,
+      "losses/sft": 0.8853281140327454,
+      "losses/total": 0.702146053314209,
+      "ref_logps/chosen": -43.87609100341797,
+      "ref_logps/rejected": -45.32105255126953,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.004192858934402466,
+      "rewards/margins": 0.0015453994274139404,
+      "rewards/rejected": 0.002647459041327238,
+      "step": 54
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 29.713237762451172,
+      "learning_rate": 2.08018154311649e-08,
+      "logps/chosen": -32.37765884399414,
+      "logps/rejected": -28.05956268310547,
+      "loss": 0.6928,
+      "losses/dpo": 0.6871882677078247,
+      "losses/sft": 0.3716886639595032,
+      "losses/total": 0.6871882677078247,
+      "ref_logps/chosen": -32.31031799316406,
+      "ref_logps/rejected": -27.984233856201172,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.006734251976013184,
+      "rewards/margins": 0.0007985057309269905,
+      "rewards/rejected": -0.007532757706940174,
+      "step": 55
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 35.79730224609375,
+      "learning_rate": 2.118003025718608e-08,
+      "logps/chosen": -40.47246551513672,
+      "logps/rejected": -19.13518524169922,
+      "loss": 0.6856,
+      "losses/dpo": 0.7039137482643127,
+      "losses/sft": 1.24458646774292,
+      "losses/total": 0.7039137482643127,
+      "ref_logps/chosen": -40.5584831237793,
+      "ref_logps/rejected": -19.068458557128906,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.008601206354796886,
+      "rewards/margins": 0.015273911878466606,
+      "rewards/rejected": -0.006672704126685858,
+      "step": 56
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 32.738399505615234,
+      "learning_rate": 2.155824508320726e-08,
+      "logps/chosen": -29.533538818359375,
+      "logps/rejected": -25.8756160736084,
+      "loss": 0.6997,
+      "losses/dpo": 0.6963517665863037,
+      "losses/sft": 0.3461652100086212,
+      "losses/total": 0.6963517665863037,
+      "ref_logps/chosen": -29.44887924194336,
+      "ref_logps/rejected": -25.920303344726562,
+      "rewards/accuracies": 0.125,
+      "rewards/chosen": -0.008466106839478016,
+      "rewards/margins": -0.012934833765029907,
+      "rewards/rejected": 0.004468727391213179,
+      "step": 57
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 40.13890838623047,
+      "learning_rate": 2.193645990922844e-08,
+      "logps/chosen": -46.109703063964844,
+      "logps/rejected": -19.18845558166504,
+      "loss": 0.6912,
+      "losses/dpo": 0.7002992630004883,
+      "losses/sft": 0.526166558265686,
+      "losses/total": 0.7002992630004883,
+      "ref_logps/chosen": -46.097267150878906,
+      "ref_logps/rejected": -19.13555145263672,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.0012434842064976692,
+      "rewards/margins": 0.004046953283250332,
+      "rewards/rejected": -0.005290437024086714,
+      "step": 58
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 38.2598762512207,
+      "learning_rate": 2.231467473524962e-08,
+      "logps/chosen": -23.814125061035156,
+      "logps/rejected": -24.001953125,
+      "loss": 0.6929,
+      "losses/dpo": 0.6859763264656067,
+      "losses/sft": 0.252949059009552,
+      "losses/total": 0.6859763264656067,
+      "ref_logps/chosen": -23.771499633789062,
+      "ref_logps/rejected": -23.951963424682617,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.004262452945113182,
+      "rewards/margins": 0.0007364752236753702,
+      "rewards/rejected": -0.004998928401619196,
+      "step": 59
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 23.8754825592041,
+      "learning_rate": 2.26928895612708e-08,
+      "logps/chosen": -17.547164916992188,
+      "logps/rejected": -14.98313045501709,
+      "loss": 0.6923,
+      "losses/dpo": 0.6888842582702637,
+      "losses/sft": 1.3874595165252686,
+      "losses/total": 0.6888842582702637,
+      "ref_logps/chosen": -17.561710357666016,
+      "ref_logps/rejected": -14.980500221252441,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.00145473494194448,
+      "rewards/margins": 0.0017177729168906808,
+      "rewards/rejected": -0.0002630382077768445,
+      "step": 60
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 36.69063186645508,
+      "learning_rate": 2.307110438729198e-08,
+      "logps/chosen": -25.101774215698242,
+      "logps/rejected": -42.1680908203125,
+      "loss": 0.6993,
+      "losses/dpo": 0.7086414098739624,
+      "losses/sft": 0.5335144400596619,
+      "losses/total": 0.7086414098739624,
+      "ref_logps/chosen": -25.029869079589844,
+      "ref_logps/rejected": -42.215362548828125,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.0071905553340911865,
+      "rewards/margins": -0.011917639523744583,
+      "rewards/rejected": 0.004727083723992109,
+      "step": 61
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 39.590003967285156,
+      "learning_rate": 2.344931921331316e-08,
+      "logps/chosen": -37.42922592163086,
+      "logps/rejected": -26.50375747680664,
+      "loss": 0.7016,
+      "losses/dpo": 0.6914690732955933,
+      "losses/sft": 0.6034911870956421,
+      "losses/total": 0.6914690732955933,
+      "ref_logps/chosen": -37.32251739501953,
+      "ref_logps/rejected": -26.564237594604492,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.010670709423720837,
+      "rewards/margins": -0.016718948259949684,
+      "rewards/rejected": 0.006048237904906273,
+      "step": 62
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 48.58049392700195,
+      "learning_rate": 2.382753403933434e-08,
+      "logps/chosen": -39.96108627319336,
+      "logps/rejected": -34.021141052246094,
+      "loss": 0.6989,
+      "losses/dpo": 0.688057541847229,
+      "losses/sft": 0.3272539973258972,
+      "losses/total": 0.688057541847229,
+      "ref_logps/chosen": -39.899322509765625,
+      "ref_logps/rejected": -34.07249069213867,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.0061765252612531185,
+      "rewards/margins": -0.011311227455735207,
+      "rewards/rejected": 0.0051347012631595135,
+      "step": 63
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 30.101633071899414,
+      "learning_rate": 2.420574886535552e-08,
+      "logps/chosen": -25.143138885498047,
+      "logps/rejected": -32.24797821044922,
+      "loss": 0.6911,
+      "losses/dpo": 0.7050281167030334,
+      "losses/sft": 0.7794158458709717,
+      "losses/total": 0.7050281167030334,
+      "ref_logps/chosen": -25.154491424560547,
+      "ref_logps/rejected": -32.21677017211914,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.001135277678258717,
+      "rewards/margins": 0.0042562964372336864,
+      "rewards/rejected": -0.0031210186425596476,
+      "step": 64
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 44.64031219482422,
+      "learning_rate": 2.4583963691376702e-08,
+      "logps/chosen": -49.361900329589844,
+      "logps/rejected": -26.884492874145508,
+      "loss": 0.7003,
+      "losses/dpo": 0.6936798095703125,
+      "losses/sft": 0.42234453558921814,
+      "losses/total": 0.6936798095703125,
+      "ref_logps/chosen": -49.26100158691406,
+      "ref_logps/rejected": -26.92618751525879,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": -0.010090004652738571,
+      "rewards/margins": -0.014259660616517067,
+      "rewards/rejected": 0.004169655032455921,
+      "step": 65
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 41.64508819580078,
+      "learning_rate": 2.4962178517397882e-08,
+      "logps/chosen": -47.77102279663086,
+      "logps/rejected": -48.46343231201172,
+      "loss": 0.6946,
+      "losses/dpo": 0.6899067163467407,
+      "losses/sft": 1.0442003011703491,
+      "losses/total": 0.6899067163467407,
+      "ref_logps/chosen": -47.75519943237305,
+      "ref_logps/rejected": -48.47092819213867,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.0015824679285287857,
+      "rewards/margins": -0.0023321392945945263,
+      "rewards/rejected": 0.0007496715988963842,
+      "step": 66
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 36.368980407714844,
+      "learning_rate": 2.5340393343419063e-08,
+      "logps/chosen": -28.49197769165039,
+      "logps/rejected": -41.9534912109375,
+      "loss": 0.6894,
+      "losses/dpo": 0.6909430623054504,
+      "losses/sft": 1.0093305110931396,
+      "losses/total": 0.6909430623054504,
+      "ref_logps/chosen": -28.488712310791016,
+      "ref_logps/rejected": -41.87274932861328,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.0003263830440118909,
+      "rewards/margins": 0.00774770975112915,
+      "rewards/rejected": -0.008074093610048294,
+      "step": 67
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 45.25604248046875,
+      "learning_rate": 2.5718608169440243e-08,
+      "logps/chosen": -57.19928741455078,
+      "logps/rejected": -37.23521423339844,
+      "loss": 0.6945,
+      "losses/dpo": 0.7012407779693604,
+      "losses/sft": 1.0595611333847046,
+      "losses/total": 0.7012407779693604,
+      "ref_logps/chosen": -57.22715377807617,
+      "ref_logps/rejected": -37.288795471191406,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.002786970231682062,
+      "rewards/margins": -0.002571570686995983,
+      "rewards/rejected": 0.005358540453016758,
+      "step": 68
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 27.695709228515625,
+      "learning_rate": 2.6096822995461424e-08,
+      "logps/chosen": -25.917194366455078,
+      "logps/rejected": -20.409564971923828,
+      "loss": 0.691,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.5559080839157104,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -25.978334426879883,
+      "ref_logps/rejected": -20.42633628845215,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.006114008836448193,
+      "rewards/margins": 0.00443677045404911,
+      "rewards/rejected": 0.0016772389644756913,
+      "step": 69
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 40.685455322265625,
+      "learning_rate": 2.6475037821482598e-08,
+      "logps/chosen": -28.94492530822754,
+      "logps/rejected": -28.425073623657227,
+      "loss": 0.6947,
+      "losses/dpo": 0.697110652923584,
+      "losses/sft": 0.3066958785057068,
+      "losses/total": 0.697110652923584,
+      "ref_logps/chosen": -28.83943748474121,
+      "ref_logps/rejected": -28.349702835083008,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.010548913851380348,
+      "rewards/margins": -0.00301172723993659,
+      "rewards/rejected": -0.007537186145782471,
+      "step": 70
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 47.70412826538086,
+      "learning_rate": 2.6853252647503778e-08,
+      "logps/chosen": -57.458221435546875,
+      "logps/rejected": -31.11433219909668,
+      "loss": 0.6903,
+      "losses/dpo": 0.6971299648284912,
+      "losses/sft": 1.9130865335464478,
+      "losses/total": 0.6971299648284912,
+      "ref_logps/chosen": -57.52789306640625,
+      "ref_logps/rejected": -31.124435424804688,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.006967080757021904,
+      "rewards/margins": 0.005956435110419989,
+      "rewards/rejected": 0.0010106442496180534,
+      "step": 71
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 38.20296859741211,
+      "learning_rate": 2.723146747352496e-08,
+      "logps/chosen": -48.005828857421875,
+      "logps/rejected": -41.52708435058594,
+      "loss": 0.6958,
+      "losses/dpo": 0.6851578950881958,
+      "losses/sft": 0.9613469243049622,
+      "losses/total": 0.6851578950881958,
+      "ref_logps/chosen": -47.931182861328125,
+      "ref_logps/rejected": -41.50524139404297,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.007464814465492964,
+      "rewards/margins": -0.005280792713165283,
+      "rewards/rejected": -0.0021840212866663933,
+      "step": 72
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 41.06236267089844,
+      "learning_rate": 2.760968229954614e-08,
+      "logps/chosen": -40.61887741088867,
+      "logps/rejected": -51.89995193481445,
+      "loss": 0.6866,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.6842697858810425,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -40.7034912109375,
+      "ref_logps/rejected": -51.85079574584961,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.008461451157927513,
+      "rewards/margins": 0.0133773572742939,
+      "rewards/rejected": -0.004915904253721237,
+      "step": 73
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 36.2785530090332,
+      "learning_rate": 2.798789712556732e-08,
+      "logps/chosen": -34.483856201171875,
+      "logps/rejected": -48.59144592285156,
+      "loss": 0.6861,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.28226807713508606,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -34.56922912597656,
+      "ref_logps/rejected": -48.53504943847656,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.008537418209016323,
+      "rewards/margins": 0.01417688187211752,
+      "rewards/rejected": -0.005639463663101196,
+      "step": 74
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 39.71546936035156,
+      "learning_rate": 2.83661119515885e-08,
+      "logps/chosen": -36.84790802001953,
+      "logps/rejected": -31.745813369750977,
+      "loss": 0.699,
+      "losses/dpo": 0.7019625306129456,
+      "losses/sft": 0.5089595913887024,
+      "losses/total": 0.7019625306129456,
+      "ref_logps/chosen": -36.72783660888672,
+      "ref_logps/rejected": -31.74201011657715,
+      "rewards/accuracies": 0.125,
+      "rewards/chosen": -0.012007355690002441,
+      "rewards/margins": -0.01162691693753004,
+      "rewards/rejected": -0.0003804388688877225,
+      "step": 75
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 40.46381378173828,
+      "learning_rate": 2.874432677760968e-08,
+      "logps/chosen": -30.391878128051758,
+      "logps/rejected": -32.681922912597656,
+      "loss": 0.6902,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.33321213722229004,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -30.440570831298828,
+      "ref_logps/rejected": -32.669700622558594,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.004869234748184681,
+      "rewards/margins": 0.00609170226380229,
+      "rewards/rejected": -0.001222467515617609,
+      "step": 76
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 35.94179916381836,
+      "learning_rate": 2.912254160363086e-08,
+      "logps/chosen": -31.62662124633789,
+      "logps/rejected": -18.103084564208984,
+      "loss": 0.6932,
+      "losses/dpo": 0.6998027563095093,
+      "losses/sft": 0.35783714056015015,
+      "losses/total": 0.6998027563095093,
+      "ref_logps/chosen": -31.659603118896484,
+      "ref_logps/rejected": -18.136844635009766,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.0032977815717458725,
+      "rewards/margins": -7.820117752999067e-05,
+      "rewards/rejected": 0.0033759830985218287,
+      "step": 77
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 40.7310791015625,
+      "learning_rate": 2.9500756429652042e-08,
+      "logps/chosen": -38.77708435058594,
+      "logps/rejected": -42.00521469116211,
+      "loss": 0.6934,
+      "losses/dpo": 0.694230854511261,
+      "losses/sft": 1.1035815477371216,
+      "losses/total": 0.694230854511261,
+      "ref_logps/chosen": -38.83358383178711,
+      "ref_logps/rejected": -42.06493377685547,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.005649971775710583,
+      "rewards/margins": -0.00032196100801229477,
+      "rewards/rejected": 0.0059719327837228775,
+      "step": 78
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 34.13336944580078,
+      "learning_rate": 2.987897125567322e-08,
+      "logps/chosen": -29.524137496948242,
+      "logps/rejected": -25.135303497314453,
+      "loss": 0.6985,
+      "losses/dpo": 0.7016960382461548,
+      "losses/sft": 1.4326092004776,
+      "losses/total": 0.7016960382461548,
+      "ref_logps/chosen": -29.412860870361328,
+      "ref_logps/rejected": -25.128196716308594,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.011127782054245472,
+      "rewards/margins": -0.010416990146040916,
+      "rewards/rejected": -0.0007107912097126245,
+      "step": 79
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 33.36854934692383,
+      "learning_rate": 3.02571860816944e-08,
+      "logps/chosen": -54.98564529418945,
+      "logps/rejected": -42.637054443359375,
+      "loss": 0.6894,
+      "losses/dpo": 0.6928367018699646,
+      "losses/sft": 1.0636935234069824,
+      "losses/total": 0.6928367018699646,
+      "ref_logps/chosen": -54.917259216308594,
+      "ref_logps/rejected": -42.492713928222656,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -0.006838513538241386,
+      "rewards/margins": 0.007595384027808905,
+      "rewards/rejected": -0.014433898031711578,
+      "step": 80
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 46.175270080566406,
+      "learning_rate": 3.063540090771558e-08,
+      "logps/chosen": -20.512210845947266,
+      "logps/rejected": -59.815433502197266,
+      "loss": 0.6933,
+      "losses/dpo": 0.6928080320358276,
+      "losses/sft": 0.8254087567329407,
+      "losses/total": 0.6928080320358276,
+      "ref_logps/chosen": -20.49310874938965,
+      "ref_logps/rejected": -59.79820251464844,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -0.0019101622747257352,
+      "rewards/margins": -0.00018673017621040344,
+      "rewards/rejected": -0.0017234322149306536,
+      "step": 81
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 27.743846893310547,
+      "learning_rate": 3.101361573373676e-08,
+      "logps/chosen": -36.26102828979492,
+      "logps/rejected": -29.011253356933594,
+      "loss": 0.686,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.43656912446022034,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -36.35939407348633,
+      "ref_logps/rejected": -28.96433448791504,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.009836399927735329,
+      "rewards/margins": 0.014528321102261543,
+      "rewards/rejected": -0.0046919225715100765,
+      "step": 82
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 40.736968994140625,
+      "learning_rate": 3.1391830559757944e-08,
+      "logps/chosen": -34.923274993896484,
+      "logps/rejected": -27.28143310546875,
+      "loss": 0.6902,
+      "losses/dpo": 0.692177414894104,
+      "losses/sft": 0.6479670405387878,
+      "losses/total": 0.692177414894104,
+      "ref_logps/chosen": -34.980430603027344,
+      "ref_logps/rejected": -27.278949737548828,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.005715560168027878,
+      "rewards/margins": 0.0059640295803546906,
+      "rewards/rejected": -0.0002484679571352899,
+      "step": 83
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 56.229461669921875,
+      "learning_rate": 3.177004538577912e-08,
+      "logps/chosen": -75.5541763305664,
+      "logps/rejected": -44.489219665527344,
+      "loss": 0.6972,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.37301287055015564,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -75.5169906616211,
+      "ref_logps/rejected": -44.53234100341797,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.003718042280524969,
+      "rewards/margins": -0.008030176162719727,
+      "rewards/rejected": 0.004312134347856045,
+      "step": 84
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 52.807533264160156,
+      "learning_rate": 3.2148260211800305e-08,
+      "logps/chosen": -43.525230407714844,
+      "logps/rejected": -53.981651306152344,
+      "loss": 0.6941,
+      "losses/dpo": 0.6933290362358093,
+      "losses/sft": 0.3423389196395874,
+      "losses/total": 0.6933290362358093,
+      "ref_logps/chosen": -43.519981384277344,
+      "ref_logps/rejected": -53.99076843261719,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0005246158689260483,
+      "rewards/margins": -0.0014358097687363625,
+      "rewards/rejected": 0.0009111943654716015,
+      "step": 85
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 42.125762939453125,
+      "learning_rate": 3.252647503782148e-08,
+      "logps/chosen": -39.07624053955078,
+      "logps/rejected": -25.18871307373047,
+      "loss": 0.6888,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.278859406709671,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -39.1097412109375,
+      "ref_logps/rejected": -25.13461685180664,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.003349810838699341,
+      "rewards/margins": 0.008759424090385437,
+      "rewards/rejected": -0.0054096137173473835,
+      "step": 86
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 55.95197296142578,
+      "learning_rate": 3.2904689863842666e-08,
+      "logps/chosen": -77.34832763671875,
+      "logps/rejected": -34.583526611328125,
+      "loss": 0.6954,
+      "losses/dpo": 0.7006169557571411,
+      "losses/sft": 0.2824787497520447,
+      "losses/total": 0.7006169557571411,
+      "ref_logps/chosen": -77.23515319824219,
+      "ref_logps/rejected": -34.51284408569336,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.011317819356918335,
+      "rewards/margins": -0.004249686375260353,
+      "rewards/rejected": -0.007068133447319269,
+      "step": 87
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 52.838645935058594,
+      "learning_rate": 3.328290468986384e-08,
+      "logps/chosen": -52.83473205566406,
+      "logps/rejected": -52.93191909790039,
+      "loss": 0.6906,
+      "losses/dpo": 0.6846374273300171,
+      "losses/sft": 0.8209627270698547,
+      "losses/total": 0.6846374273300171,
+      "ref_logps/chosen": -52.82656478881836,
+      "ref_logps/rejected": -52.87086486816406,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.0008166790939867496,
+      "rewards/margins": 0.0052884165197610855,
+      "rewards/rejected": -0.006105095613747835,
+      "step": 88
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 23.7989559173584,
+      "learning_rate": 3.366111951588503e-08,
+      "logps/chosen": -18.322547912597656,
+      "logps/rejected": -22.132925033569336,
+      "loss": 0.688,
+      "losses/dpo": 0.6949195265769958,
+      "losses/sft": 0.16063320636749268,
+      "losses/total": 0.6949195265769958,
+      "ref_logps/chosen": -18.328706741333008,
+      "ref_logps/rejected": -22.033899307250977,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0006158978212624788,
+      "rewards/margins": 0.010518422350287437,
+      "rewards/rejected": -0.009902525693178177,
+      "step": 89
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 50.99845886230469,
+      "learning_rate": 3.40393343419062e-08,
+      "logps/chosen": -36.93525314331055,
+      "logps/rejected": -53.35734176635742,
+      "loss": 0.6886,
+      "losses/dpo": 0.6997432112693787,
+      "losses/sft": 0.27666959166526794,
+      "losses/total": 0.6997432112693787,
+      "ref_logps/chosen": -36.983482360839844,
+      "ref_logps/rejected": -53.31299591064453,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.004823225550353527,
+      "rewards/margins": 0.009258443489670753,
+      "rewards/rejected": -0.004435217939317226,
+      "step": 90
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 36.005638122558594,
+      "learning_rate": 3.441754916792738e-08,
+      "logps/chosen": -36.347496032714844,
+      "logps/rejected": -38.60151672363281,
+      "loss": 0.6834,
+      "losses/dpo": 0.7013605833053589,
+      "losses/sft": 0.44756028056144714,
+      "losses/total": 0.7013605833053589,
+      "ref_logps/chosen": -36.41640853881836,
+      "ref_logps/rejected": -38.47201919555664,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.006891525350511074,
+      "rewards/margins": 0.01984090730547905,
+      "rewards/rejected": -0.01294938288629055,
+      "step": 91
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 39.57761764526367,
+      "learning_rate": 3.479576399394856e-08,
+      "logps/chosen": -38.53449249267578,
+      "logps/rejected": -34.8283805847168,
+      "loss": 0.6852,
+      "losses/dpo": 0.6864519715309143,
+      "losses/sft": 0.46143075823783875,
+      "losses/total": 0.6864519715309143,
+      "ref_logps/chosen": -38.64277648925781,
+      "ref_logps/rejected": -34.775604248046875,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.010828555561602116,
+      "rewards/margins": 0.016106223687529564,
+      "rewards/rejected": -0.0052776699885725975,
+      "step": 92
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 42.923465728759766,
+      "learning_rate": 3.5173978819969736e-08,
+      "logps/chosen": -42.63794708251953,
+      "logps/rejected": -40.95237350463867,
+      "loss": 0.691,
+      "losses/dpo": 0.6687886714935303,
+      "losses/sft": 0.9120340943336487,
+      "losses/total": 0.6687886714935303,
+      "ref_logps/chosen": -42.673095703125,
+      "ref_logps/rejected": -40.939796447753906,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": 0.0035148384049534798,
+      "rewards/margins": 0.004772840533405542,
+      "rewards/rejected": -0.0012580035254359245,
+      "step": 93
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 35.36449432373047,
+      "learning_rate": 3.555219364599092e-08,
+      "logps/chosen": -30.112043380737305,
+      "logps/rejected": -36.460514068603516,
+      "loss": 0.6971,
+      "losses/dpo": 0.6995805501937866,
+      "losses/sft": 0.8373420834541321,
+      "losses/total": 0.6995805501937866,
+      "ref_logps/chosen": -30.007692337036133,
+      "ref_logps/rejected": -36.43312454223633,
+      "rewards/accuracies": 0.125,
+      "rewards/chosen": -0.010435272008180618,
+      "rewards/margins": -0.007696258835494518,
+      "rewards/rejected": -0.0027390120085328817,
+      "step": 94
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 50.16935348510742,
+      "learning_rate": 3.59304084720121e-08,
+      "logps/chosen": -43.58863067626953,
+      "logps/rejected": -39.95378875732422,
+      "loss": 0.697,
+      "losses/dpo": 0.6925528645515442,
+      "losses/sft": 0.41814520955085754,
+      "losses/total": 0.6925528645515442,
+      "ref_logps/chosen": -43.46757888793945,
+      "ref_logps/rejected": -39.90763854980469,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.012105096131563187,
+      "rewards/margins": -0.0074903108179569244,
+      "rewards/rejected": -0.004614785313606262,
+      "step": 95
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 46.554473876953125,
+      "learning_rate": 3.630862329803328e-08,
+      "logps/chosen": -42.712406158447266,
+      "logps/rejected": -38.224639892578125,
+      "loss": 0.6944,
+      "losses/dpo": 0.6929863691329956,
+      "losses/sft": 0.6638107895851135,
+      "losses/total": 0.6929863691329956,
+      "ref_logps/chosen": -42.69837188720703,
+      "ref_logps/rejected": -38.235469818115234,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.001403355854563415,
+      "rewards/margins": -0.002486205194145441,
+      "rewards/rejected": 0.0010828491067513824,
+      "step": 96
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 50.87691879272461,
+      "learning_rate": 3.668683812405446e-08,
+      "logps/chosen": -47.13976287841797,
+      "logps/rejected": -52.6763916015625,
+      "loss": 0.6873,
+      "losses/dpo": 0.7065359354019165,
+      "losses/sft": 1.3687249422073364,
+      "losses/total": 0.7065359354019165,
+      "ref_logps/chosen": -47.076499938964844,
+      "ref_logps/rejected": -52.49290466308594,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.006325757130980492,
+      "rewards/margins": 0.012022912502288818,
+      "rewards/rejected": -0.01834867149591446,
+      "step": 97
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 28.458797454833984,
+      "learning_rate": 3.706505295007564e-08,
+      "logps/chosen": -20.34185791015625,
+      "logps/rejected": -19.174640655517578,
+      "loss": 0.6927,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.6108443140983582,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -20.374011993408203,
+      "ref_logps/rejected": -19.19711685180664,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0032154680229723454,
+      "rewards/margins": 0.0009677950292825699,
+      "rewards/rejected": 0.0022476734593510628,
+      "step": 98
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 38.541439056396484,
+      "learning_rate": 3.744326777609682e-08,
+      "logps/chosen": -35.34051513671875,
+      "logps/rejected": -39.98661422729492,
+      "loss": 0.6994,
+      "losses/dpo": 0.6870009303092957,
+      "losses/sft": 1.0313451290130615,
+      "losses/total": 0.6870009303092957,
+      "ref_logps/chosen": -35.31980895996094,
+      "ref_logps/rejected": -40.09027099609375,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.002070808317512274,
+      "rewards/margins": -0.012436461634933949,
+      "rewards/rejected": 0.010365653783082962,
+      "step": 99
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 32.7611198425293,
+      "learning_rate": 3.7821482602118e-08,
+      "logps/chosen": -28.146602630615234,
+      "logps/rejected": -32.26946258544922,
+      "loss": 0.691,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.7257211804389954,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -28.161846160888672,
+      "ref_logps/rejected": -32.24116134643555,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0015242458321154118,
+      "rewards/margins": 0.004354441538453102,
+      "rewards/rejected": -0.0028301957063376904,
+      "step": 100
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 48.886348724365234,
+      "learning_rate": 3.819969742813918e-08,
+      "logps/chosen": -24.085975646972656,
+      "logps/rejected": -53.43666458129883,
+      "loss": 0.6898,
+      "losses/dpo": 0.6925390958786011,
+      "losses/sft": 0.7717922329902649,
+      "losses/total": 0.6925390958786011,
+      "ref_logps/chosen": -24.032161712646484,
+      "ref_logps/rejected": -53.31513977050781,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.005381548311561346,
+      "rewards/margins": 0.006771028973162174,
+      "rewards/rejected": -0.012152576819062233,
+      "step": 101
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 37.729129791259766,
+      "learning_rate": 3.8577912254160364e-08,
+      "logps/chosen": -39.28417205810547,
+      "logps/rejected": -33.66284942626953,
+      "loss": 0.6891,
+      "losses/dpo": 0.6890721321105957,
+      "losses/sft": 0.11002220213413239,
+      "losses/total": 0.6890721321105957,
+      "ref_logps/chosen": -39.35163879394531,
+      "ref_logps/rejected": -33.64738464355469,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.006746518891304731,
+      "rewards/margins": 0.008293158374726772,
+      "rewards/rejected": -0.0015466390177607536,
+      "step": 102
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 36.61566162109375,
+      "learning_rate": 3.895612708018154e-08,
+      "logps/chosen": -32.80045700073242,
+      "logps/rejected": -26.297595977783203,
+      "loss": 0.6938,
+      "losses/dpo": 0.6875132918357849,
+      "losses/sft": 0.7562694549560547,
+      "losses/total": 0.6875132918357849,
+      "ref_logps/chosen": -32.78963088989258,
+      "ref_logps/rejected": -26.29842185974121,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.0010827300138771534,
+      "rewards/margins": -0.0011653571855276823,
+      "rewards/rejected": 8.262717165052891e-05,
+      "step": 103
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 35.90760803222656,
+      "learning_rate": 3.9334341906202725e-08,
+      "logps/chosen": -25.600452423095703,
+      "logps/rejected": -26.01952362060547,
+      "loss": 0.6905,
+      "losses/dpo": 0.6928426623344421,
+      "losses/sft": 0.560149610042572,
+      "losses/total": 0.6928426623344421,
+      "ref_logps/chosen": -25.59532928466797,
+      "ref_logps/rejected": -25.960851669311523,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -0.000512230210006237,
+      "rewards/margins": 0.0053548929281532764,
+      "rewards/rejected": -0.005867123603820801,
+      "step": 104
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 41.926326751708984,
+      "learning_rate": 3.97125567322239e-08,
+      "logps/chosen": -24.64590835571289,
+      "logps/rejected": -47.35856628417969,
+      "loss": 0.6903,
+      "losses/dpo": 0.6918473243713379,
+      "losses/sft": 0.6405824422836304,
+      "losses/total": 0.6918473243713379,
+      "ref_logps/chosen": -24.60011863708496,
+      "ref_logps/rejected": -47.25457000732422,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.004578936379402876,
+      "rewards/margins": 0.0058210729621350765,
+      "rewards/rejected": -0.010400008410215378,
+      "step": 105
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 44.33354568481445,
+      "learning_rate": 4.0090771558245085e-08,
+      "logps/chosen": -32.457977294921875,
+      "logps/rejected": -37.6816291809082,
+      "loss": 0.7012,
+      "losses/dpo": 0.6904045343399048,
+      "losses/sft": 0.4094051718711853,
+      "losses/total": 0.6904045343399048,
+      "ref_logps/chosen": -32.348350524902344,
+      "ref_logps/rejected": -37.73137283325195,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.010962343774735928,
+      "rewards/margins": -0.015936829149723053,
+      "rewards/rejected": 0.004974484443664551,
+      "step": 106
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 50.068389892578125,
+      "learning_rate": 4.046898638426626e-08,
+      "logps/chosen": -57.488861083984375,
+      "logps/rejected": -45.16341018676758,
+      "loss": 0.7013,
+      "losses/dpo": 0.6909089684486389,
+      "losses/sft": 0.7873000502586365,
+      "losses/total": 0.6909089684486389,
+      "ref_logps/chosen": -57.47578811645508,
+      "ref_logps/rejected": -45.31059265136719,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.0013074756134301424,
+      "rewards/margins": -0.01602541282773018,
+      "rewards/rejected": 0.014717936515808105,
+      "step": 107
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 61.65685272216797,
+      "learning_rate": 4.084720121028744e-08,
+      "logps/chosen": -76.06755065917969,
+      "logps/rejected": -79.65231323242188,
+      "loss": 0.6854,
+      "losses/dpo": 0.7170514464378357,
+      "losses/sft": 1.8121122121810913,
+      "losses/total": 0.7170514464378357,
+      "ref_logps/chosen": -76.13177490234375,
+      "ref_logps/rejected": -79.55784606933594,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.006422186270356178,
+      "rewards/margins": 0.01586904563009739,
+      "rewards/rejected": -0.009446859359741211,
+      "step": 108
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 41.34520721435547,
+      "learning_rate": 4.1225416036308624e-08,
+      "logps/chosen": -30.94908332824707,
+      "logps/rejected": -31.076162338256836,
+      "loss": 0.7003,
+      "losses/dpo": 0.7072453498840332,
+      "losses/sft": 0.6860259175300598,
+      "losses/total": 0.7072453498840332,
+      "ref_logps/chosen": -30.907779693603516,
+      "ref_logps/rejected": -31.175491333007812,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.004130268469452858,
+      "rewards/margins": -0.014063073322176933,
+      "rewards/rejected": 0.009932804852724075,
+      "step": 109
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 39.601680755615234,
+      "learning_rate": 4.16036308623298e-08,
+      "logps/chosen": -29.889883041381836,
+      "logps/rejected": -29.010711669921875,
+      "loss": 0.6875,
+      "losses/dpo": 0.6812564730644226,
+      "losses/sft": 0.33825215697288513,
+      "losses/total": 0.6812564730644226,
+      "ref_logps/chosen": -29.933313369750977,
+      "ref_logps/rejected": -28.940982818603516,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.004343009553849697,
+      "rewards/margins": 0.011315608397126198,
+      "rewards/rejected": -0.006972598843276501,
+      "step": 110
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 43.49833297729492,
+      "learning_rate": 4.1981845688350985e-08,
+      "logps/chosen": -28.440502166748047,
+      "logps/rejected": -31.334300994873047,
+      "loss": 0.6842,
+      "losses/dpo": 0.6787906885147095,
+      "losses/sft": 0.6913360953330994,
+      "losses/total": 0.6787906885147095,
+      "ref_logps/chosen": -28.545825958251953,
+      "ref_logps/rejected": -31.25735092163086,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.01053229533135891,
+      "rewards/margins": 0.018227243795990944,
+      "rewards/rejected": -0.007694947998970747,
+      "step": 111
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 32.026695251464844,
+      "learning_rate": 4.236006051437216e-08,
+      "logps/chosen": -37.945472717285156,
+      "logps/rejected": -24.111562728881836,
+      "loss": 0.6889,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.2837979197502136,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -37.992347717285156,
+      "ref_logps/rejected": -24.073345184326172,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.004687649197876453,
+      "rewards/margins": 0.008509386330842972,
+      "rewards/rejected": -0.003821736667305231,
+      "step": 112
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 39.62443161010742,
+      "learning_rate": 4.2738275340393346e-08,
+      "logps/chosen": -37.079071044921875,
+      "logps/rejected": -39.647586822509766,
+      "loss": 0.6918,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.15460079908370972,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -37.13287353515625,
+      "ref_logps/rejected": -39.6738395690918,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.005380382761359215,
+      "rewards/margins": 0.002755367662757635,
+      "rewards/rejected": 0.002625015564262867,
+      "step": 113
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 38.18842697143555,
+      "learning_rate": 4.311649016641452e-08,
+      "logps/chosen": -25.791282653808594,
+      "logps/rejected": -32.9259033203125,
+      "loss": 0.6933,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.6581792831420898,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -25.775999069213867,
+      "ref_logps/rejected": -32.91184997558594,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.0015283345710486174,
+      "rewards/margins": -0.00012282142415642738,
+      "rewards/rejected": -0.0014055129140615463,
+      "step": 114
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 52.07583999633789,
+      "learning_rate": 4.34947049924357e-08,
+      "logps/chosen": -41.8395881652832,
+      "logps/rejected": -44.14722442626953,
+      "loss": 0.6899,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.8708814382553101,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -41.82201385498047,
+      "ref_logps/rejected": -44.06287384033203,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.001757478341460228,
+      "rewards/margins": 0.006677555851638317,
+      "rewards/rejected": -0.00843503512442112,
+      "step": 115
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 33.54158020019531,
+      "learning_rate": 4.387291981845688e-08,
+      "logps/chosen": -32.49979782104492,
+      "logps/rejected": -17.52083969116211,
+      "loss": 0.6952,
+      "losses/dpo": 0.6925131678581238,
+      "losses/sft": 0.14384517073631287,
+      "losses/total": 0.6925131678581238,
+      "ref_logps/chosen": -32.42995071411133,
+      "ref_logps/rejected": -17.491504669189453,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.006984531879425049,
+      "rewards/margins": -0.0040509942919015884,
+      "rewards/rejected": -0.0029335380531847477,
+      "step": 116
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 37.52485275268555,
+      "learning_rate": 4.425113464447806e-08,
+      "logps/chosen": -39.653236389160156,
+      "logps/rejected": -26.307018280029297,
+      "loss": 0.6949,
+      "losses/dpo": 0.6901547908782959,
+      "losses/sft": 0.5280264019966125,
+      "losses/total": 0.6901547908782959,
+      "ref_logps/chosen": -39.55583953857422,
+      "ref_logps/rejected": -26.243160247802734,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.009739398956298828,
+      "rewards/margins": -0.003353619948029518,
+      "rewards/rejected": -0.006385779473930597,
+      "step": 117
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 44.086185455322266,
+      "learning_rate": 4.462934947049924e-08,
+      "logps/chosen": -39.232688903808594,
+      "logps/rejected": -61.158592224121094,
+      "loss": 0.6861,
+      "losses/dpo": 0.6927624940872192,
+      "losses/sft": 0.3766418695449829,
+      "losses/total": 0.6927624940872192,
+      "ref_logps/chosen": -39.291038513183594,
+      "ref_logps/rejected": -61.07466125488281,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.005835331045091152,
+      "rewards/margins": 0.014228129759430885,
+      "rewards/rejected": -0.008392797783017159,
+      "step": 118
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 35.55986785888672,
+      "learning_rate": 4.500756429652042e-08,
+      "logps/chosen": -26.11540985107422,
+      "logps/rejected": -20.935359954833984,
+      "loss": 0.698,
+      "losses/dpo": 0.6918513774871826,
+      "losses/sft": 1.0432565212249756,
+      "losses/total": 0.6918513774871826,
+      "ref_logps/chosen": -26.03144073486328,
+      "ref_logps/rejected": -20.945743560791016,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.008397170342504978,
+      "rewards/margins": -0.009435638785362244,
+      "rewards/rejected": 0.0010384678607806563,
+      "step": 119
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 39.22114944458008,
+      "learning_rate": 4.53857791225416e-08,
+      "logps/chosen": -29.81409454345703,
+      "logps/rejected": -39.67371368408203,
+      "loss": 0.6915,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.7685847878456116,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -29.86721420288086,
+      "ref_logps/rejected": -39.69112777709961,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.005311918444931507,
+      "rewards/margins": 0.003570293541997671,
+      "rewards/rejected": 0.0017416239716112614,
+      "step": 120
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 34.359615325927734,
+      "learning_rate": 4.576399394856278e-08,
+      "logps/chosen": -17.609037399291992,
+      "logps/rejected": -33.15533447265625,
+      "loss": 0.6903,
+      "losses/dpo": 0.6952323913574219,
+      "losses/sft": 0.41854506731033325,
+      "losses/total": 0.6952323913574219,
+      "ref_logps/chosen": -17.5953369140625,
+      "ref_logps/rejected": -33.08323669433594,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0013699710834771395,
+      "rewards/margins": 0.005839664954692125,
+      "rewards/rejected": -0.007209635339677334,
+      "step": 121
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 36.893653869628906,
+      "learning_rate": 4.614220877458396e-08,
+      "logps/chosen": -29.187511444091797,
+      "logps/rejected": -28.67523956298828,
+      "loss": 0.6968,
+      "losses/dpo": 0.6963827013969421,
+      "losses/sft": 0.2885913848876953,
+      "losses/total": 0.6963827013969421,
+      "ref_logps/chosen": -29.182682037353516,
+      "ref_logps/rejected": -28.74280548095703,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.0004826546646654606,
+      "rewards/margins": -0.007239234633743763,
+      "rewards/rejected": 0.00675658043473959,
+      "step": 122
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 30.719209671020508,
+      "learning_rate": 4.652042360060514e-08,
+      "logps/chosen": -26.74323081970215,
+      "logps/rejected": -24.072521209716797,
+      "loss": 0.6907,
+      "losses/dpo": 0.6998178958892822,
+      "losses/sft": 0.9175503849983215,
+      "losses/total": 0.6998178958892822,
+      "ref_logps/chosen": -26.759307861328125,
+      "ref_logps/rejected": -24.038970947265625,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.001607751939445734,
+      "rewards/margins": 0.004962801933288574,
+      "rewards/rejected": -0.00335504999384284,
+      "step": 123
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 35.995609283447266,
+      "learning_rate": 4.689863842662632e-08,
+      "logps/chosen": -24.474266052246094,
+      "logps/rejected": -30.566516876220703,
+      "loss": 0.6852,
+      "losses/dpo": 0.7060790061950684,
+      "losses/sft": 0.5684544444084167,
+      "losses/total": 0.7060790061950684,
+      "ref_logps/chosen": -24.575531005859375,
+      "ref_logps/rejected": -30.50520133972168,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.010126292705535889,
+      "rewards/margins": 0.016257895156741142,
+      "rewards/rejected": -0.006131601519882679,
+      "step": 124
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 40.390480041503906,
+      "learning_rate": 4.72768532526475e-08,
+      "logps/chosen": -47.56864547729492,
+      "logps/rejected": -27.925689697265625,
+      "loss": 0.6911,
+      "losses/dpo": 0.6941109299659729,
+      "losses/sft": 0.761809766292572,
+      "losses/total": 0.6941109299659729,
+      "ref_logps/chosen": -47.580238342285156,
+      "ref_logps/rejected": -27.895416259765625,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.0011593580711632967,
+      "rewards/margins": 0.004186606500297785,
+      "rewards/rejected": -0.0030272486619651318,
+      "step": 125
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 43.70969772338867,
+      "learning_rate": 4.765506807866868e-08,
+      "logps/chosen": -47.28679656982422,
+      "logps/rejected": -31.254356384277344,
+      "loss": 0.6898,
+      "losses/dpo": 0.701967179775238,
+      "losses/sft": 0.725429356098175,
+      "losses/total": 0.701967179775238,
+      "ref_logps/chosen": -47.3563346862793,
+      "ref_logps/rejected": -31.255340576171875,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.006953556090593338,
+      "rewards/margins": 0.006855214014649391,
+      "rewards/rejected": 9.834184311330318e-05,
+      "step": 126
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 32.73550033569336,
+      "learning_rate": 4.803328290468986e-08,
+      "logps/chosen": -29.04664421081543,
+      "logps/rejected": -38.83638000488281,
+      "loss": 0.6903,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.23836013674736023,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -29.14601707458496,
+      "ref_logps/rejected": -38.87748718261719,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.009937328286468983,
+      "rewards/margins": 0.005826891399919987,
+      "rewards/rejected": 0.004110437817871571,
+      "step": 127
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 33.31587219238281,
+      "learning_rate": 4.841149773071104e-08,
+      "logps/chosen": -17.702585220336914,
+      "logps/rejected": -33.756351470947266,
+      "loss": 0.6883,
+      "losses/dpo": 0.6917657852172852,
+      "losses/sft": 0.695366621017456,
+      "losses/total": 0.6917657852172852,
+      "ref_logps/chosen": -17.81790542602539,
+      "ref_logps/rejected": -33.7726936340332,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.011532002128660679,
+      "rewards/margins": 0.009898019023239613,
+      "rewards/rejected": 0.0016339838039129972,
+      "step": 128
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 34.78561782836914,
+      "learning_rate": 4.878971255673222e-08,
+      "logps/chosen": -31.822887420654297,
+      "logps/rejected": -31.502132415771484,
+      "loss": 0.6918,
+      "losses/dpo": 0.6878577470779419,
+      "losses/sft": 0.6958499550819397,
+      "losses/total": 0.6878577470779419,
+      "ref_logps/chosen": -31.875709533691406,
+      "ref_logps/rejected": -31.527292251586914,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.005282568745315075,
+      "rewards/margins": 0.002766538178548217,
+      "rewards/rejected": 0.0025160307995975018,
+      "step": 129
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 39.473045349121094,
+      "learning_rate": 4.9167927382753404e-08,
+      "logps/chosen": -48.424564361572266,
+      "logps/rejected": -37.70363998413086,
+      "loss": 0.6907,
+      "losses/dpo": 0.6958696246147156,
+      "losses/sft": 1.0931024551391602,
+      "losses/total": 0.6958696246147156,
+      "ref_logps/chosen": -48.39546585083008,
+      "ref_logps/rejected": -37.62480545043945,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.0029095769859850407,
+      "rewards/margins": 0.004973782226443291,
+      "rewards/rejected": -0.007883358746767044,
+      "step": 130
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 40.11540985107422,
+      "learning_rate": 4.954614220877458e-08,
+      "logps/chosen": -47.275970458984375,
+      "logps/rejected": -24.532180786132812,
+      "loss": 0.6891,
+      "losses/dpo": 0.686172366142273,
+      "losses/sft": 0.23291173577308655,
+      "losses/total": 0.686172366142273,
+      "ref_logps/chosen": -47.30372619628906,
+      "ref_logps/rejected": -24.47891616821289,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0027754190377891064,
+      "rewards/margins": 0.008101915940642357,
+      "rewards/rejected": -0.005326497368514538,
+      "step": 131
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 35.008811950683594,
+      "learning_rate": 4.9924357034795765e-08,
+      "logps/chosen": -22.004154205322266,
+      "logps/rejected": -25.471179962158203,
+      "loss": 0.6983,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.5767898559570312,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -22.0076847076416,
+      "ref_logps/rejected": -25.575210571289062,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.00035313377156853676,
+      "rewards/margins": -0.010050022043287754,
+      "rewards/rejected": 0.010403156280517578,
+      "step": 132
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 48.81768035888672,
+      "learning_rate": 5.030257186081694e-08,
+      "logps/chosen": -41.83135223388672,
+      "logps/rejected": -61.698612213134766,
+      "loss": 0.6887,
+      "losses/dpo": 0.69705730676651,
+      "losses/sft": 0.6419839262962341,
+      "losses/total": 0.69705730676651,
+      "ref_logps/chosen": -41.8665771484375,
+      "ref_logps/rejected": -61.644554138183594,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.0035225865431129932,
+      "rewards/margins": 0.008928824216127396,
+      "rewards/rejected": -0.005406237207353115,
+      "step": 133
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 35.35521697998047,
+      "learning_rate": 5.0680786686838126e-08,
+      "logps/chosen": -42.63499450683594,
+      "logps/rejected": -18.00177001953125,
+      "loss": 0.6887,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.22498027980327606,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -42.69081115722656,
+      "ref_logps/rejected": -17.965286254882812,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": 0.005581945180892944,
+      "rewards/margins": 0.009230315685272217,
+      "rewards/rejected": -0.0036483705043792725,
+      "step": 134
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 37.32074737548828,
+      "learning_rate": 5.10590015128593e-08,
+      "logps/chosen": -28.98600196838379,
+      "logps/rejected": -28.591854095458984,
+      "loss": 0.6916,
+      "losses/dpo": 0.6911870241165161,
+      "losses/sft": 0.32320740818977356,
+      "losses/total": 0.6911870241165161,
+      "ref_logps/chosen": -28.984033584594727,
+      "ref_logps/rejected": -28.557567596435547,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.00019681453704833984,
+      "rewards/margins": 0.003232082584872842,
+      "rewards/rejected": -0.0034288971219211817,
+      "step": 135
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 38.866493225097656,
+      "learning_rate": 5.143721633888049e-08,
+      "logps/chosen": -40.32871627807617,
+      "logps/rejected": -39.726768493652344,
+      "loss": 0.69,
+      "losses/dpo": 0.6905131936073303,
+      "losses/sft": 1.158380150794983,
+      "losses/total": 0.6905131936073303,
+      "ref_logps/chosen": -40.46543884277344,
+      "ref_logps/rejected": -39.79895782470703,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.013672376051545143,
+      "rewards/margins": 0.0064532398246228695,
+      "rewards/rejected": 0.007219136226922274,
+      "step": 136
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 37.07709884643555,
+      "learning_rate": 5.1815431164901664e-08,
+      "logps/chosen": -27.539257049560547,
+      "logps/rejected": -27.673938751220703,
+      "loss": 0.7006,
+      "losses/dpo": 0.6891195774078369,
+      "losses/sft": 1.0018857717514038,
+      "losses/total": 0.6891195774078369,
+      "ref_logps/chosen": -27.427831649780273,
+      "ref_logps/rejected": -27.70747947692871,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.011142684146761894,
+      "rewards/margins": -0.014496791176497936,
+      "rewards/rejected": 0.0033541086595505476,
+      "step": 137
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 38.496559143066406,
+      "learning_rate": 5.219364599092285e-08,
+      "logps/chosen": -22.926864624023438,
+      "logps/rejected": -38.9444580078125,
+      "loss": 0.699,
+      "losses/dpo": 0.6990444660186768,
+      "losses/sft": 0.39025476574897766,
+      "losses/total": 0.6990444660186768,
+      "ref_logps/chosen": -22.85403823852539,
+      "ref_logps/rejected": -38.987823486328125,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.007282680831849575,
+      "rewards/margins": -0.011619050055742264,
+      "rewards/rejected": 0.004336369223892689,
+      "step": 138
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 22.79326629638672,
+      "learning_rate": 5.2571860816944025e-08,
+      "logps/chosen": -18.254146575927734,
+      "logps/rejected": -24.10297966003418,
+      "loss": 0.69,
+      "losses/dpo": 0.6989715099334717,
+      "losses/sft": 0.46928051114082336,
+      "losses/total": 0.6989715099334717,
+      "ref_logps/chosen": -18.28525161743164,
+      "ref_logps/rejected": -24.071096420288086,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.003110540099442005,
+      "rewards/margins": 0.006298840045928955,
+      "rewards/rejected": -0.0031883001793175936,
+      "step": 139
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 45.97643280029297,
+      "learning_rate": 5.2950075642965196e-08,
+      "logps/chosen": -33.01849365234375,
+      "logps/rejected": -40.95310974121094,
+      "loss": 0.698,
+      "losses/dpo": 0.6955220103263855,
+      "losses/sft": 0.3267292082309723,
+      "losses/total": 0.6955220103263855,
+      "ref_logps/chosen": -32.911224365234375,
+      "ref_logps/rejected": -40.94056701660156,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.010726600885391235,
+      "rewards/margins": -0.00947202555835247,
+      "rewards/rejected": -0.0012545764911919832,
+      "step": 140
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 35.01676940917969,
+      "learning_rate": 5.332829046898638e-08,
+      "logps/chosen": -37.70044708251953,
+      "logps/rejected": -25.797130584716797,
+      "loss": 0.6923,
+      "losses/dpo": 0.6944743990898132,
+      "losses/sft": 0.46885228157043457,
+      "losses/total": 0.6944743990898132,
+      "ref_logps/chosen": -37.66633605957031,
+      "ref_logps/rejected": -25.7462158203125,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.0034110904671251774,
+      "rewards/margins": 0.0016802309546619654,
+      "rewards/rejected": -0.005091322120279074,
+      "step": 141
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 46.63210678100586,
+      "learning_rate": 5.3706505295007557e-08,
+      "logps/chosen": -44.68863296508789,
+      "logps/rejected": -44.29450988769531,
+      "loss": 0.6917,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.3401167392730713,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -44.73490524291992,
+      "ref_logps/rejected": -44.3111572265625,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.004627292975783348,
+      "rewards/margins": 0.0029626486357301474,
+      "rewards/rejected": 0.0016646445728838444,
+      "step": 142
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 39.051780700683594,
+      "learning_rate": 5.408472012102874e-08,
+      "logps/chosen": -32.727230072021484,
+      "logps/rejected": -24.590736389160156,
+      "loss": 0.6913,
+      "losses/dpo": 0.6861852407455444,
+      "losses/sft": 0.829793393611908,
+      "losses/total": 0.6861852407455444,
+      "ref_logps/chosen": -32.69562911987305,
+      "ref_logps/rejected": -24.520002365112305,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.0031601781956851482,
+      "rewards/margins": 0.003913199529051781,
+      "rewards/rejected": -0.007073379121720791,
+      "step": 143
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 52.55834197998047,
+      "learning_rate": 5.446293494704992e-08,
+      "logps/chosen": -63.71332550048828,
+      "logps/rejected": -49.05284881591797,
+      "loss": 0.6943,
+      "losses/dpo": 0.6921186447143555,
+      "losses/sft": 0.3558349907398224,
+      "losses/total": 0.6921186447143555,
+      "ref_logps/chosen": -63.716209411621094,
+      "ref_logps/rejected": -49.073699951171875,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.00028821243904531,
+      "rewards/margins": -0.0017975267255678773,
+      "rewards/rejected": 0.0020857397466897964,
+      "step": 144
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 40.16358184814453,
+      "learning_rate": 5.48411497730711e-08,
+      "logps/chosen": -45.54597473144531,
+      "logps/rejected": -27.000322341918945,
+      "loss": 0.6912,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 1.1267122030258179,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -45.54884338378906,
+      "ref_logps/rejected": -26.963048934936523,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.00028649577870965004,
+      "rewards/margins": 0.00401383638381958,
+      "rewards/rejected": -0.0037273408379405737,
+      "step": 145
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 38.33798599243164,
+      "learning_rate": 5.521936459909228e-08,
+      "logps/chosen": -42.34248352050781,
+      "logps/rejected": -28.343761444091797,
+      "loss": 0.6904,
+      "losses/dpo": 0.7000496983528137,
+      "losses/sft": 1.3564696311950684,
+      "losses/total": 0.7000496983528137,
+      "ref_logps/chosen": -42.38239288330078,
+      "ref_logps/rejected": -28.327028274536133,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.003990745171904564,
+      "rewards/margins": 0.0056642405688762665,
+      "rewards/rejected": -0.0016734958626329899,
+      "step": 146
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 34.70563888549805,
+      "learning_rate": 5.559757942511346e-08,
+      "logps/chosen": -29.78840446472168,
+      "logps/rejected": -30.435606002807617,
+      "loss": 0.6914,
+      "losses/dpo": 0.6939413547515869,
+      "losses/sft": 0.7547587156295776,
+      "losses/total": 0.6939413547515869,
+      "ref_logps/chosen": -29.82432746887207,
+      "ref_logps/rejected": -30.435745239257812,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.003592371940612793,
+      "rewards/margins": 0.0035784118808805943,
+      "rewards/rejected": 1.3959361240267754e-05,
+      "step": 147
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 40.98420333862305,
+      "learning_rate": 5.597579425113464e-08,
+      "logps/chosen": -34.28458786010742,
+      "logps/rejected": -41.45768737792969,
+      "loss": 0.6991,
+      "losses/dpo": 0.6844355463981628,
+      "losses/sft": 1.0345463752746582,
+      "losses/total": 0.6844355463981628,
+      "ref_logps/chosen": -34.26361083984375,
+      "ref_logps/rejected": -41.55308151245117,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.0020977025851607323,
+      "rewards/margins": -0.01163715124130249,
+      "rewards/rejected": 0.009539449587464333,
+      "step": 148
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 32.001407623291016,
+      "learning_rate": 5.6354009077155823e-08,
+      "logps/chosen": -23.2528076171875,
+      "logps/rejected": -28.89508819580078,
+      "loss": 0.6966,
+      "losses/dpo": 0.6931471824645996,
+      "losses/sft": 0.8549447655677795,
+      "losses/total": 0.6931471824645996,
+      "ref_logps/chosen": -23.25066375732422,
+      "ref_logps/rejected": -28.96025848388672,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": -0.00021452270448207855,
+      "rewards/margins": -0.006731557194143534,
+      "rewards/rejected": 0.006517034024000168,
+      "step": 149
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 38.03684616088867,
+      "learning_rate": 5.6732223903177e-08,
+      "logps/chosen": -34.9976692199707,
+      "logps/rejected": -34.524696350097656,
+      "loss": 0.6917,
+      "losses/dpo": 0.6935884356498718,
+      "losses/sft": 1.2651445865631104,
+      "losses/total": 0.6935884356498718,
+      "ref_logps/chosen": -34.93056106567383,
+      "ref_logps/rejected": -34.42768478393555,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.006710815709084272,
+      "rewards/margins": 0.002990209963172674,
+      "rewards/rejected": -0.009701025672256947,
+      "step": 150
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 13215,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 50,
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

VideoSAVi-Qwen-7B-all-CLIP/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7e56fdac56581473a945a81ece63119f76f91d4ad99edf97676c6c0a3234bde
+size 7864

VideoSAVi-Qwen-7B-all-CLIP/zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/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 . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+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 param_shapes.items():
+
+        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_file, 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_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``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
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+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_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    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_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)