f

app.py

@@ -25,29 +25,32 @@ base_model = AutoModelForCausalLM.from_pretrained(
 model = PeftModel.from_pretrained(
     base_model,
     ADAPTER_REPO,
-    ignore_mismatched_sizes=True,  # Add this parameter
+    ignore_mismatched_sizes=True,
 )
 
-def classify_text(text):
+def classify_food(text):
     """
-    Simple prompting approach: we ask the model to return a single classification label 
-    (e.g., 'positive', 'negative', etc.).
-    You can refine this prompt, add chain-of-thought, or multiple classes as needed.
+    Classify or extract food-related terms from the input text.
     """
-    prompt = f"Below is some text.\nText: {text}\nPlease classify the sentiment (positive or negative):"
+    prompt = f"Below is some text. Please identify and classify food-related terms.\nText: {text}\nFood classification or extraction:"
     inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
     with torch.no_grad():
-        outputs = model.generate(**inputs, max_new_tokens=64)
+        outputs = model.generate(
+            **inputs,
+            max_new_tokens=64,
+            temperature=0.7,  # Adjust temperature for creativity
+            top_p=0.9,       # Adjust top_p for diversity
+        )
     answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
     return answer
 
 with gr.Blocks() as demo:
-    gr.Markdown("## Qwen + LoRA Adapter: Text Classification Demo")
-    input_box = gr.Textbox(lines=3, label="Enter text")
-    output_box = gr.Textbox(lines=3, label="Model's generated output (classification)")
+    gr.Markdown("## Qwen + LoRA Adapter: Food Classification/Extraction Demo")
+    input_box = gr.Textbox(lines=3, label="Enter text containing food items")
+    output_box = gr.Textbox(lines=3, label="Model's classification or extraction output")
 
-    classify_btn = gr.Button("Classify")
-    classify_btn.click(fn=classify_text, inputs=input_box, outputs=output_box)
+    classify_btn = gr.Button("Analyze Food Terms")
+    classify_btn.click(fn=classify_food, inputs=input_box, outputs=output_box)
 
 if __name__ == "__main__":
     demo.launch()