Update app.py

app.py

@@ -8,11 +8,13 @@ tokenizer = AutoTokenizer.from_pretrained(model_path)
 model = AutoModelForCausalLM.from_pretrained(model_path, torch_dtype=torch.bfloat16)
 
 # Define the inference function
-def generate_response(input_text, temperature, top_k, top_p):
+def generate_response(input_text, temperature, top_k, top_p, progress=gr.Progress()):
+    progress(0, "Processing input...")
     before = f"# This is the assembly code:\n"#prompt
     after = "\n# What is the source code?\n"#prompt
     input_func = before+input_text.strip()+after
     inputs = tokenizer(input_func, return_tensors="pt")
+    progress(0.3, "Running inference...")
     outputs = model.generate(
         **inputs,
         max_length=512,  # Adjust this if needed
@@ -21,7 +23,9 @@ def generate_response(input_text, temperature, top_k, top_p):
     #    top_k=int(top_k),
     #    top_p=float(top_p),
     #    temperature=float(temperature)
+    progress(0.8, "Decoding response...")
     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    progress(1, "Done!")
     # Split the response into assembly and source code (if applicable)
     if "# This is the assembly code:" in response:
         parts = response.split("# What is the source code?")
@@ -41,8 +45,8 @@ interface = gr.Interface(
         gr.Slider(0.1, 1.0, value=0.95, step=0.05, label="Top-p")
     ],
     outputs=[
-        gr.Markdown(label="Assembly Code"),
-        gr.Markdown(label="Source Code")
+        gr.Markdown(label="Assembly Code", placeholder = "Here goes the input function ...."),
+        gr.Markdown(label="Source Code", placeholder = "Here goes the enhanced function ....")
     ],
 
     title="Falcon decompiler Interactive Demo",