Updates

app.py

@@ -1,14 +1,24 @@
 import streamlit as st
 from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+from peft import PeftModel
 
 st.set_page_config(page_title="Hugging Face Chatbot", layout="centered")
-st.title("Hugging Face Chatbot")
+st.title("Hugging Face Chatbot with LoRA")
 
 @st.cache_resource
 def load_model():
-    # Load tokenizer and model from Hugging Face
-    tokenizer = AutoTokenizer.from_pretrained("Grandediw/lora_model_finetuned", use_fast=True)
-    model = AutoModelForCausalLM.from_pretrained("Grandediw/lora_model_finetuned", device_map="auto", trust_remote_code=True)
+    # Replace this with the actual base model used during LoRA fine-tuning
+    base_model_name = "unsloth/Llama-3.2-1B-Instruct"
+    
+    # Load the base model and tokenizer
+    tokenizer = AutoTokenizer.from_pretrained(base_model_name, use_fast=False)
+    base_model = AutoModelForCausalLM.from_pretrained(base_model_name, device_map="auto", trust_remote_code=True)
+    
+    # Load the LoRA adapter weights
+    # Replace "Grandediw/lora_model_finetuned" with your actual LoRA model repo
+    model = PeftModel.from_pretrained(base_model, "Grandediw/lora_model_finetuned")
+    
+    # Create a pipeline for text generation
     chat_pipeline = pipeline(
         "text-generation",
         model=model,
@@ -33,19 +43,19 @@ for message in st.session_state.messages:
 
 # User input
 if prompt := st.chat_input("Ask me anything:"):
-    # Display user message and store it
+    # Display user message
     st.chat_message("user").markdown(prompt)
     st.session_state.messages.append({"role": "user", "content": prompt})
 
     # Generate response
     with st.spinner("Thinking..."):
+        # Generate text with the pipeline
         response = chat_pipeline(prompt)[0]["generated_text"]
-        # The model might return the prompt + response together.
-        # If that's the case, remove the prompt from the start.
+        # Remove the prompt from the start if it's included
         if response.startswith(prompt):
             response = response[len(prompt):].strip()
 
-    # Display and store assistant response
+    # Display assistant response
     with st.chat_message("assistant"):
         st.markdown(response)
     st.session_state.messages.append({"role": "assistant", "content": response})

requirements.txt

@@ -1,3 +1,4 @@
-streamlit
-transformers
-torch
+streamlit==1.25.0
+transformers==4.34.0
+torch==2.0.1
+peft==0.7.0