Update app.py

app.py

@@ -1,40 +1,16 @@
-import os
 import gradio as gr
 import torch
 from transformers import BertTokenizer, BertForSequenceClassification, GPT2LMHeadModel, GPT2Tokenizer
-from huggingface_hub import list_files
 
-# --- 1. SETUP AND MODEL LOADING WITH DEBUGGING ---
+# --- 1. LOAD MODELS FROM HUGGING FACE HUB ---
 
 BERT_REPO = "Sankeerth004/fitbuddy-bert-intent"
 GPT2_REPO = "Sankeerth004/fitbuddy-gpt2-chatbot"
 
-print("--- Starting Application ---")
-print("--- Running Pre-flight Checks for Model Files ---")
-
-# --- BERT Model File Check ---
-try:
-    print(f"Checking files for BERT repo: {BERT_REPO}")
-    bert_files = list_files(repo_id=BERT_REPO)
-    print(f"Found BERT files: {bert_files}")
-    
-    # Check for the most important tokenizer file
-    if 'vocab.txt' not in [os.path.basename(f) for f in bert_files]:
-        raise FileNotFoundError(
-            "CRITICAL ERROR: 'vocab.txt' is missing from the BERT repository. "
-            "Please upload all files from your local './models/bert_intent_model/' folder."
-        )
-    print("BERT file check successful.")
-
-except Exception as e:
-    print(f"Could not check BERT repository files. Error: {e}")
-    # Stop the app if we can't verify the files
-    raise e
-
-# --- Load Models ---
 print("Loading models from Hugging Face Hub...")
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
+# Load models
 bert_tokenizer = BertTokenizer.from_pretrained(BERT_REPO)
 bert_model = BertForSequenceClassification.from_pretrained(BERT_REPO).to(device)
 bert_model.eval()
@@ -46,15 +22,15 @@ gpt2_model.config.pad_token_id = gpt2_tokenizer.eos_token_id
 
 print("Models loaded successfully!")
 
-# --- 2. DEFINE HELPER FUNCTIONS (No Changes Needed Here) ---
+# --- 2. DEFINE HELPER FUNCTIONS ---
 
 intent_mapping = {
     0: "Body Building", 1: "Meal Plan Recommendation", 2: "Recommend Meditation or Yoga",
     3: "Suggest Recovery Exercises", 4: "Weight Loss"
 }
 
-def predict_intent(input_text):
-    inputs = bert_tokenizer(input_text, return_tensors='pt', padding=True, truncation=True, max_length=512).to(device)
+def predict_intent(text):
+    inputs = bert_tokenizer(text, return_tensors='pt', padding=True, truncation=True, max_length=512).to(device)
     with torch.no_grad():
         outputs = bert_model(**inputs)
     return intent_mapping.get(torch.argmax(outputs.logits, dim=1).item(), "Unknown Intent")
@@ -69,7 +45,7 @@ def generate_response(prompt, max_length=200):
     )
     return gpt2_tokenizer.decode(output[0], skip_special_tokens=True).split('[A]')[-1].strip()
 
-# --- 3. CREATE THE GRADIO CHATBOT FUNCTION (No Changes Needed Here) ---
+# --- 3. CREATE THE GRADIO CHATBOT FUNCTION ---
 
 def chat_with_fitbuddy(user_input, history):
     if not user_input.strip(): return "Please enter a valid question."
@@ -77,7 +53,7 @@ def chat_with_fitbuddy(user_input, history):
     prompt = f"[Q] {user_input} [Intent: {intent}]"
     return generate_response(prompt)
 
-# --- 4. LAUNCH THE GRADIO INTERFACE (No Changes Needed Here) ---
+# --- 4. LAUNCH THE GRADIO INTERFACE ---
 
 iface = gr.ChatInterface(
     fn=chat_with_fitbuddy,