update app.py & requirements

app.py

@@ -2,30 +2,28 @@ import os
 import torch
 import gradio as gr
 from peft import PeftModel, PeftConfig
-from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
+from transformers import AutoModelForCausalLM, AutoTokenizer
 
 
 def load_model(model_path):
     config = PeftConfig.from_pretrained(model_path)
 
-    bnb_config = BitsAndBytesConfig(
-        load_in_4bit=True,
-        bnb_4bit_quant_type="nf4",
-        bnb_4bit_compute_dtype=torch.float16,
-    )
-
-    model = AutoModelForCausalLM.from_pretrained(
+    base_model = AutoModelForCausalLM.from_pretrained(
         config.base_model_name_or_path,
-        quantization_config=bnb_config,
         trust_remote_code=True,
-        token=os.environ["HF_TOKEN"],
+        # token=os.environ["HF_TOKEN"],
     )
-    model.config.use_cache = False
+
+    base_model.config.use_cache = False
     tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)
     tokenizer.pad_token = tokenizer.eos_token
 
     # Load the Lora model
-    model = PeftModel.from_pretrained(model, model_path)
+    model = PeftModel.from_pretrained(
+        base_model, 
+        model_path,
+        low_cpu_mem_usage=True
+        )
 
     return model, tokenizer
 

requirements.txt

@@ -1,5 +1,4 @@
 torch
 transformers
 peft
-gradio
-bitsandbytes
+gradio