Update app.py

app.py

@@ -1,9 +1,14 @@
 import gradio as gr
 from llama_cpp import Llama
+import os
 
-# Provide the relative path to the first shard of your model
+# Path to the first shard of the model
 model_path = "DeepSeek-R1-Zero-Q4_K_M/DeepSeek-R1-Zero-Q4_K_M-00001-of-00009.gguf"
 
+# Debugging: Verify working directory and model path
+print("Current working directory:", os.getcwd())
+print("Full model path:", os.path.join(os.getcwd(), model_path))
+
 # Initialize the model
 try:
     model = Llama(model_path=model_path, n_threads=8)
@@ -16,7 +21,8 @@ def predict(prompt):
     try:
         # Generate output using the model
         output = model(prompt)
-        return output["choices"][0]["text"]  # Adjust this based on output structure
+        # Extract and return the text from the response
+        return output["choices"][0]["text"]
     except Exception as e:
         return f"Error during inference: {e}"