major change, used biomed_llama 7b

app.py

@@ -192,12 +192,19 @@ def initialize_model():
 def initialize_llm():
     try:
         print("Starting LLM initialization...")
+        # Add quantization config
+        quantization_config = BitsAndBytesConfig(
+            load_in_4bit=True,
+            bnb_4bit_compute_dtype=torch.float16
+        )
+        
         model = AutoModel.from_pretrained(
             "ContactDoctor/Bio-Medical-MultiModal-Llama-3-8B-V1",
             device_map="auto",
             torch_dtype=torch.float16,
             trust_remote_code=True,
-            low_cpu_mem_usage=True
+            low_cpu_mem_usage=True,
+            quantization_config=quantization_config
         )
         print("Model loaded successfully")
         

requirements.txt

@@ -11,8 +11,8 @@ timm==0.4.12
 numpy==1.26.4
 einops==0.8.0
 fvcore==0.1.5.post20221221
-transformers==4.34.0
-sentencepiece==0.1.99
+# transformers==4.34.0
+sentencepiece
 ftfy==6.1.1
 regex==2023.10.3
 nltk==3.8.1
@@ -25,7 +25,8 @@ pycocotools==2.0.7
 diffdist==0.1
 scikit-image==0.21.0
 mup==1.0.0
-accelerate==0.23.0
+accelerate
+bitsandbytes
 kornia==0.7.0
 infinibatch==0.1.1
 open-clip-torch==2.26.1
@@ -55,6 +56,7 @@ tokenizers==0.14.1
 # 
 # torch>=2.0.0
 # transformers>=4.34.0
+ transformers==4.36.0
 # gradio>=4.40.0
 # Pillow>=9.0.0
 # numpy>=1.21.0