Update app.py

app.py

@@ -1,5 +1,5 @@
 import os
-import requests
+from transformers import pipeline
 from sentence_transformers import SentenceTransformer
 from sklearn.metrics.pairwise import cosine_similarity
 import numpy as np
@@ -15,24 +15,16 @@ df = pd.read_csv('combined_questions_and_answers.csv')
 # Encode all questions in the dataset
 question_embeddings = model.encode(df['Question'].tolist())
 
-# Hugging Face API details for Meta-Llama 70B
-API_URL = "https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B"
-headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_KEY')}"}
+# Hugging Face API details for Meta-Llama 3B
+pipe = pipeline("text-generation", model="meta-llama/Meta-Llama-3-8B-Instruct")
 
-# Function to call Hugging Face API to refine and translate text
+# Function to refine and translate text using Meta-Llama 3B
 def refine_text(prompt):
-    payload = {
-        "inputs": prompt,
-        "parameters": {
-            "max_new_tokens": 800,
-            "temperature": 0.7
-        }
-    }
-    response = requests.post(API_URL, headers=headers, json=payload)
-    response_json = response.json()
-    if isinstance(response_json, list) and len(response_json) > 0:
-        return response_json[0].get('generated_text', '')
-    return "Error in refining text."
+    messages = [
+        {"role": "user", "content": prompt},
+    ]
+    response = pipe(messages)
+    return response[0]['generated_text']
 
 # Function to find the most similar question and provide the answer
 def get_answer(user_question, threshold=0.30):
@@ -49,7 +41,7 @@ def get_answer(user_question, threshold=0.30):
         similar_question_idx = np.argmax(similarities)
         # Retrieve the corresponding answer
         answer = df.iloc[similar_question_idx]['Answer']
-        # Refine the answer using Meta-Llama 70B
+        # Refine the answer using Meta-Llama 3B
         refined_answer = refine_text(f"Refine this answer: {answer}")
         return refined_answer, max_similarity
     else: