Update app.py

app.py

@@ -8,13 +8,15 @@ import transformers
 from transformers import AutoTokenizer, AutoModelForCausalLM
 
 
-offload_folder = '/path/to/your/offload/folder'
-model = YourModel(..., offload_folder=offload_folder)
+offload_folder = 'C:\model_weights'
+
+
 model = AutoModelForCausalLM.from_pretrained("PyaeSoneK/LlamaV2LegalFineTuned",
                                     device_map='auto',
                                   torch_dtype=torch.float16,
                                   use_auth_token= st.secrets['hf_access_token'],
-                                    offload_folder=offload_folder,)
+                                    offload_folder=offload_folder,
+                                    low_cpu_mem_usage=True)
                                             #  load_in_4bit=True
 
 tokenizer = AutoTokenizer.from_pretrained("PyaeSoneK/LlamaV2LegalFineTuned",