Create app.py

app.py

@@ -4,7 +4,7 @@ import torch
 import gradio as gr
 import os
 
-# Authenticate with Hugging Face
+# Authenticate
 login(token=os.environ.get("HF_TOKEN"))
 
 # Configuration
@@ -18,38 +18,40 @@ quant_config = BitsAndBytesConfig(
     bnb_4bit_quant_type="nf4"
 )
 
-# Load model with optimizations
-tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR)
-model = AutoModelForCausalLM.from_pretrained(
-    MODEL_NAME,
-    quantization_config=quant_config,
-    device_map="auto",
-    torch_dtype=torch.float16,
-    cache_dir=CACHE_DIR
-)
+# Load model
+try:
+    tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR)
+    model = AutoModelForCausalLM.from_pretrained(
+        MODEL_NAME,
+        quantization_config=quant_config,
+        device_map="auto",
+        torch_dtype=torch.float16,
+        cache_dir=CACHE_DIR
+    )
+except Exception as e:
+    raise gr.Error(f"Model loading failed: {str(e)}")
 
 def solve_math(question):
-    prompt = f"""Solve this math problem step by step:
-    
-Question: {question}
-Answer:"""
-    
-    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
-    outputs = model.generate(
-        **inputs,
-        max_new_tokens=256,
-        temperature=0.7,
-        do_sample=True
-    )
-    return tokenizer.decode(outputs[0], skip_special_tokens=True)
+    try:
+        prompt = f"Solve step by step: {question}\nAnswer:"
+        inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+        outputs = model.generate(
+            **inputs,
+            max_new_tokens=256,
+            temperature=0.3,
+            do_sample=True
+        )
+        return tokenizer.decode(outputs[0], skip_special_tokens=True)
+    except Exception as e:
+        return f"Error: {str(e)}"
 
 # Gradio Interface
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown("""<h1><center>🧮 Gemma-2B Math Solver</center></h1>""")
     with gr.Row():
         question = gr.Textbox(
-            label="Enter your math problem",
-            placeholder="e.g., What is the derivative of x^2?",
+            label="Math Problem",
+            placeholder="What is 2^10 + 5*3?",
             lines=3
         )
     with gr.Row():
@@ -70,6 +72,5 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
 if __name__ == "__main__":
     demo.launch(
         server_name="0.0.0.0",
-        server_port=7860,
-        share=False
+        server_port=7860
     )