update read model

InfiMed.py

@@ -195,7 +195,7 @@ class InfiMed(PreTrainedModel):
                 raise NotImplementedError(
                     f'{config.llm_config.architectures[0]} is not implemented.')
     
-        self.tokenizer = AutoTokenizer.from_pretrained(config.llm_config._name_or_path, use_fast=True)
+        self.tokenizer = AutoTokenizer.from_pretrained(config._name_or_path, use_fast=True)
         self.tokenizer.add_special_tokens({"additional_special_tokens": ["<|endofchunk|>", "<s>", "<|pad|>"]})
         self.tokenizer.pad_token = "<|pad|>"
         self.tokenizer.bos_token = "<s>"
@@ -218,7 +218,7 @@ class InfiMed(PreTrainedModel):
         self.img_context_token_id = 151655
 
         self.image_processor = SiglipImageProcessor.from_pretrained(
-            config.vision_config._name_or_path,
+            config._name_or_path,
             size={"height": 384, "width": 384},  
             resample=PIL.Image.Resampling.BICUBIC,  
             crop_size={"height": 384, "width": 384},  

config.json

@@ -1,71 +1,3 @@
-{
-  "architectures": [
-    "InfiMed"
-  ],
-  "llm_config": {
-    "_name_or_path": ".",
-    "architectures": [
-      "Qwen3ForCausalLM"
-    ],
-    "pad_token_id": 151670,
-    "bos_token_id": 128245,
-    "eos_token_id": 151645,
-    "attention_bias": false,
-    "attention_dropout": 0.0,
-    "head_dim": 128,
-    "hidden_act": "silu",
-    "hidden_size": 2560,
-    "initializer_range": 0.02,
-    "intermediate_size": 9728,
-    "max_position_embeddings": 40960,
-    "max_window_layers": 36,
-    "model_type": "qwen3",
-    "num_attention_heads": 32,
-    "num_hidden_layers": 36,
-    "num_key_value_heads": 8,
-    "rms_norm_eps": 1e-06,
-    "rope_scaling": null,
-    "rope_theta": 1000000,
-    "sliding_window": null,
-    "torch_dtype": "bfloat16",
-    "use_cache": true,
-    "use_sliding_window": false,
-    "vocab_size": 151936,
-    "tie_word_embeddings": false
-  },
-  "load_precision": "bf16",
-  "max_length": 32,
-  "repetition_penalty": 1.2,
-  "run_dir": "qwen3-4b-instruct-continue-training+stage-finetune+x7",
-  "seed": 7,
-  "stage": "finetune",
-  "temperature": 0.0,
-  "top_p": 0.0001,
-  "torch_dtype": "bfloat16",
-  "trackers": [
-    "jsonl",
-    "wandb"
-  ],
-  "transformers_version": "4.52.4",
-  "vision_config": {
-    "_name_or_path": ".",
-    "architectures": [
-      "SiglipModel"
-    ],
-    "hidden_act": "gelu",
-    "attention_dropout": 0.0,
-    "hidden_size": 1152,
-    "image_size": 384,
-    "initializer_factor": 1.0,
-    "intermediate_size": 4304,
-    "layer_norm_eps": 1e-06,
-    "model_type": "siglip_vision_model",
-    "num_attention_heads": 16,
-    "num_channels": 3,
-    "num_hidden_layers": 27,
-    "patch_size": 14,
-    "torch_dtype": "bfloat16"
-  },
-  "wandb_entity": null,
-  "wandb_project": "mmpretrain"
-}
+version https://git-lfs.github.com/spec/v1
+oid sha256:01a7c1602b680b7301e429c8e2006000308bdc2f34e79d4cad87c4ddeb0076b6
+size 1675