Update app.py

app.py

@@ -1,9 +1,16 @@
+import os
+
+# Disable Gradio SSR (better for Render)
+os.environ["GRADIO_SSR_MODE"] = "False"
+
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 import torch
 
+
 MODEL_NAME = "gaussalgo/T5-LM-Large-text2sql-spider"
 
+
 tokenizer = None
 model = None
 
@@ -15,6 +22,7 @@ def load_model():
     global tokenizer, model
 
     if model is None:
+
         print("Loading model...")
 
         tokenizer = AutoTokenizer.from_pretrained(
@@ -27,71 +35,106 @@ def load_model():
 
         model.to(device)
 
+        model.eval()
+
         print(f"Model ready on {device}")
 
 
+
 def generate_sql(question, context):
 
-    load_model()
+    try:
 
-    input_text = f"{question} | {context}"
+        load_model()
 
-    inputs = tokenizer(
-        input_text,
-        return_tensors="pt",
-        max_length=512,
-        truncation=True
-    ).to(device)
 
+        input_text = f"{question} | {context}"
 
-    outputs = model.generate(
-        **inputs,
-        max_new_tokens=128,
-        num_beams=4,
-        early_stopping=True
-    )
 
+        inputs = tokenizer(
+            input_text,
+            return_tensors="pt",
+            max_length=512,
+            truncation=True
+        ).to(device)
+
+
+        with torch.no_grad():
+
+            outputs = model.generate(
+                **inputs,
+                max_new_tokens=128,
+                num_beams=4,
+                early_stopping=True
+            )
+
+
+        sql = tokenizer.decode(
+            outputs[0],
+            skip_special_tokens=True
+        )
+
+
+        return sql
 
-    sql = tokenizer.decode(
-        outputs[0],
-        skip_special_tokens=True
-    )
 
-    return sql
+    except Exception as e:
+
+        return f"Error: {str(e)}"
+
 
 
 
 with gr.Blocks() as demo:
 
-    gr.Markdown("# NL2SQL API")
+
+    gr.Markdown(
+        "# NL2SQL API\nGenerate SQL from Natural Language"
+    )
+
 
     with gr.Row():
 
         question = gr.Textbox(
-            label="Question"
+            label="Question",
+            placeholder="Example: Find all users"
         )
 
+
         context = gr.Textbox(
-            label="Context"
+            label="Database Schema",
+            placeholder="Example: users(id,name,email)"
         )
 
 
     output = gr.Textbox(
-        label="SQL"
+        label="Generated SQL",
+        lines=5
     )
 
 
-    btn = gr.Button("Submit")
+    btn = gr.Button(
+        "Generate SQL"
+    )
 
 
     btn.click(
         fn=generate_sql,
-        inputs=[question, context],
+        inputs=[
+            question,
+            context
+        ],
         outputs=output
     )
 
 
-demo.launch(
-    server_name="0.0.0.0",
-    server_port=7860
-)
+
+if __name__ == "__main__":
+
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=int(
+            os.environ.get("PORT",7860)
+        ),
+        share=False
+    )