Update app.py

app.py

@@ -1,23 +1,43 @@
 import gradio as gr
-from llama_cpp import Llama
 from huggingface_hub import hf_hub_download
+from llama_cpp import Llama
 
-
-# Download model from HF Hub to local cache
+# Download the model from Hugging Face Hub (only first time, then cached)
+print("Downloading model...")
 model_path = hf_hub_download(
-    repo_id="unsloth/gemma-3-4b-it-GGUF",
-    filename="gemma-3-4b-it.Q4_K_M.gguf"   # or correct filename
+    repo_id="TheBloke/gemma-7b-it-GGUF",
+    filename="gemma-7b-it.Q4_K_M.gguf"
 )
+print("Model downloaded to:", model_path)
 
-# Load model
-llm = Llama(model_path=model_path, n_ctx=2048)
-
+# Load the model with llama-cpp-python
+print("Loading model...")
+llm = Llama(
+    model_path=model_path,
+    n_ctx=2048  # You can adjust this as per your memory limits
+)
+print("Model loaded.")
 
-# Chat function
+# Define a function to interact with the model
 def chat(prompt):
-    output = llm(prompt, max_tokens=512)
-    return output['choices'][0]['text']
+    print(f"User input: {prompt}")
+    output = llm(
+        prompt,
+        max_tokens=512,
+        stop=["</s>", "User:", "Assistant:"]
+    )
+    reply = output['choices'][0]['text'].strip()
+    print(f"Model reply: {reply}")
+    return reply
+
+# Create a simple Gradio UI
+iface = gr.Interface(
+    fn=chat,
+    inputs="text",
+    outputs="text",
+    title="Gemma 7B IT Chatbot",
+    description="Running a GGUF model (Gemma 7B IT) on Hugging Face Space using llama-cpp-python"
+)
 
-# Gradio UI
-iface = gr.Interface(fn=chat, inputs="text", outputs="text")
+# Launch the app
 iface.launch()