Upload model

CXR_LLAVA_HF.py

@@ -7,6 +7,7 @@ from functools import partial
 from transformers import TextIteratorStreamer
 from transformers import StoppingCriteria, GenerationConfig
 from threading import Thread
+from dataclasses import dataclass
 
 # Model Constants
 IGNORE_INDEX = -100
@@ -19,6 +20,11 @@ class AttrDict(dict):
     def __init__(self, *args, **kwargs):
         super(AttrDict, self).__init__(*args, **kwargs)
         self.__dict__ = self
+    def __getattr__(self, key):
+        if key in self:
+            return self[key]
+        raise AttributeError(f"'AttrDict' object has no attribute '{key}'")
+
 
 class CXRLLAVAConfig(PretrainedConfig):
     model_type = "CXR-LLAVA"
@@ -45,7 +51,6 @@ class CXRLLAVAModel(PreTrainedModel):
         self.tokenizer.cls_token = self.tokenizer.unk_token
         self.tokenizer.mask_token = self.tokenizer.unk_token
 
-        from open_clip.model import CLIPVisionCfg
         vision_cfg = CLIPVisionCfg(**config.clip_vision_cfg)
 
         self.generation_config =  GenerationConfig.from_pretrained(config._name_or_path)
@@ -669,3 +674,32 @@ class KeywordsStoppingCriteria(StoppingCriteria):
             if keyword in outputs:
                 return True
         return False
+@dataclass
+class CLIPVisionCfg:
+    layers: Union[Tuple[int, int, int, int], int] = 12
+    width: int = 768
+    head_width: int = 64
+    mlp_ratio: float = 4.0
+    patch_size: int = 16
+    image_size: Union[Tuple[int, int], int] = 224
+
+    ls_init_value: Optional[float] = None  # layer scale initial value
+    patch_dropout: float = 0.  # what fraction of patches to dropout during training (0 would mean disabled and no patches dropped) - 0.5 to 0.75 recommended in the paper for optimal results
+    attentional_pool: bool = False  # whether to use attentional pooler in the last embedding layer (overrides pool_type)
+    attn_pooler_queries: int = 256  # n_queries for attentional pooler
+    attn_pooler_heads: int = 8  # n heads for attentional_pooling
+    no_ln_pre: bool = False  # disable pre transformer LayerNorm
+    pos_embed_type: str = 'learnable'
+    final_ln_after_pool: bool = False  # apply final LayerNorm after pooling
+    pool_type: str = 'tok'
+    output_tokens: bool = False
+    act_kwargs: Optional[dict] = None
+    norm_kwargs: Optional[dict] = None
+
+    timm_model_name: Optional[str] = None  # a valid model name overrides layers, width, patch_size
+    timm_model_pretrained: bool = False  # use (imagenet) pretrained weights for named model
+    timm_pool: str = 'avg'  # feature pooling for timm model ('abs_attn', 'rot_attn', 'avg', '')
+    timm_proj: str = 'linear'  # linear projection for timm model output ('linear', 'mlp', '')
+    timm_proj_bias: bool = False  # enable bias final projection
+    timm_drop: float = 0.  # head dropout
+    timm_drop_path: Optional[float] = None  # backbone stochastic depth

config.json

@@ -1,8 +1,9 @@
 {
-  "_name_or_path": "CXR-LLAVA-v2",
+  "_name_or_path": "G:\\Temp\\finetune_result\\LLAMA2-7B-CHAT_ViT-L-16-512_MOREKEYWORD_LN_PATCH_FINETUNE_ChexpertJSON_POSTTRAIN_25000_DIST",
   "architectures": [
     "CXRLLAVAModel"
   ],
+  "attn_implementation": null,
   "auto_map": {
     "AutoConfig": "CXR_LLAVA_HF.CXRLLAVAConfig",
     "AutoModel": "CXR_LLAVA_HF.CXRLLAVAModel"
@@ -25,7 +26,7 @@
     "std": 0.3821719215686275
   },
   "llama": {
-    "_name_or_path": "CXR-LLAVA-v2",
+    "_name_or_path": "/home/jovyan/llava/SW_LLAVA/LLAMA2-7B-CHAT_ViT-L-16-512_MOREKEYWORD_LN_PATCH_FINETUNE_ChexpertJSON_POSTTRAIN",
     "add_cross_attention": false,
     "architectures": [
       "LlamaForCausalLM"
@@ -104,7 +105,7 @@
     "vocab_size": 32000
   },
   "llama_model_dtype": "bf16",
-  "llama_model_path": "CXR-LLAVA-v2",
+  "llama_model_path": "/home/jovyan/llava/SW_LLAVA/LLAMA2-7B-CHAT_ViT-L-16-512_MOREKEYWORD_LN_PATCH_FINETUNE_ChexpertJSON_POSTTRAIN",
   "mm_projector_dim": 1024,
   "mm_projector_dtype": "fp32",
   "mm_projector_path": null,
@@ -112,7 +113,7 @@
   "mm_use_im_start_end": false,
   "model_type": "CXR-LLAVA",
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.34.0",
+  "transformers_version": "4.36.2",
   "tune_mm_mlp_adapter": false,
   "use_cache": false
 }

generation_config.json

@@ -1,4 +1,4 @@
 {
   "_from_model_config": true,
-  "transformers_version": "4.34.0"
+  "transformers_version": "4.36.2"
 }

model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a30a3337ea0e778627c417863e19621d3df385d2b843ac98c97fd23737f5053
+size 4917365472

model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f34846c3dcbf1115a4ed3992fbd3325729a3991bae7ff6418fd6161c561a604
+size 4947390880

model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42a5c74885378cc22d42cd4d2159e648fd5a13821c22a7c7daef33cdf31c53e7
+size 4228057888

model.safetensors.index.json

@@ -0,0 +1,596 @@
+{
+  "metadata": {
+    "total_size": 14092742656
+  },
+  "weight_map": {
+    "llama.embed_tokens.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.10.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.21.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.21.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.21.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.21.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.22.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.30.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.30.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.31.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "llama.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.9.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.9.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.9.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.9.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.9.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "llama.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "llama.norm.weight": "model-00003-of-00003.safetensors",
+    "lm_head.weight": "model-00001-of-00003.safetensors",
+    "mm_projector.bias": "model-00001-of-00003.safetensors",
+    "mm_projector.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.class_embedding": "model-00001-of-00003.safetensors",
+    "vision_tower.conv1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.ln_post.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.ln_post.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.ln_pre.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.ln_pre.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.positional_embedding": "model-00001-of-00003.safetensors",
+    "vision_tower.proj": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.0.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.1.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.10.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.11.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.12.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.13.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.14.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.15.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.16.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.17.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.18.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.19.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.2.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.20.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.21.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.22.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.23.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.3.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.4.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.5.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.6.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.7.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.8.mlp.c_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.attn.in_proj_bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.attn.in_proj_weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.attn.out_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.attn.out_proj.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.ln_1.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.ln_1.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.ln_2.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.ln_2.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.mlp.c_fc.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.mlp.c_fc.weight": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.mlp.c_proj.bias": "model-00001-of-00003.safetensors",
+    "vision_tower.transformer.resblocks.9.mlp.c_proj.weight": "model-00001-of-00003.safetensors"
+  }
+}