Update app.py

app.py

@@ -1,21 +1,28 @@
+# app.py
 import gradio as gr
-from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
 import torch
 
 def load_model():
-    model_name = "TheBloke/Llama-2-13B-chat-GPTQ"
-    tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
+    model_name = "meta-llama/Llama-2-13b-chat-hf"
+    
+    # Konfiguracja kwantyzacji 4-bitowej
+    quantization_config = BitsAndBytesConfig(
+        load_in_4bit=True,
+        bnb_4bit_compute_dtype=torch.float16,
+        bnb_4bit_quant_type="nf4",
+        bnb_4bit_use_double_quant=True
+    )
+    
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
     model = AutoModelForCausalLM.from_pretrained(
         model_name,
         device_map="auto",
-        trust_remote_code=True,
-        revision="main",
-        quantization_config={"load_in_4bit": True}  # Dodajemy kwantyzację 4-bitową
+        quantization_config=quantization_config,
+        trust_remote_code=True
     )
     return model, tokenizer
 
-# Reszta kodu pozostaje bez zmian
-
 def generate_response(prompt, max_length=100):
     try:
         inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
@@ -36,7 +43,6 @@ print("Ładowanie modelu...")
 model, tokenizer = load_model()
 print("Model załadowany!")
 
-# Interfejs Gradio
 iface = gr.Interface(
     fn=generate_response,
     inputs=[