Reversing Code With Alpaca Implementation

app.py

@@ -1,27 +1,30 @@
-from threading import Thread
+from threading import Thread  # Import the Thread class from the threading module
 
-import torch
-import gradio as gr
-from transformers import AutoTokenizer, AutoModel,TextIteratorStreamer
+import torch  # Import the PyTorch library
+import gradio as gr  # Import Gradio for creating a UI
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, TextIteratorStreamer  # Import Hugging Face Transformers
 
-model_id = "microsoft/codebert-base"
+# Define the Hugging Face model ID and check for available GPU (cuda)
+model_id = "declare-lab/flan-alpaca-large"
 torch_device = "cuda" if torch.cuda.is_available() else "cpu"
 print("Running on device:", torch_device)
 print("CPU threads:", torch.get_num_threads())
 
+# Load the pre-trained model based on the device
 if torch_device == "cuda":
-    model = AutoModel.from_pretrained(model_id, device_map="auto")
+    model = AutoModelForSeq2SeqLM.from_pretrained(model_id, load_in_8bit=True, device_map="auto")
 else:
-    model = AutoModel.from_pretrained(model_id)
-tokenizer = AutoTokenizer.from_pretrained(model_id)
+    model = AutoModelForSeq2SeqLM.from_pretrained(model_id)
 
+tokenizer = AutoTokenizer.from_pretrained(model_id)
 
+# Define a function to run model text generation
 def run_generation(user_text, top_p, temperature, top_k, max_new_tokens):
     # Get the model and tokenizer, and tokenize the user text.
     model_inputs = tokenizer([user_text], return_tensors="pt").to(torch_device)
 
-    # Start generation on a separate thread, so that we don't block the UI. The text is pulled from the streamer
-    # in the main thread. Adds timeout to the streamer to handle exceptions in the generation thread.
+    # Start generation on a separate thread, so that we don't block the UI.
+    # Adds timeout to the streamer to handle exceptions in the generation thread.
     streamer = TextIteratorStreamer(tokenizer, timeout=10., skip_prompt=True, skip_special_tokens=True)
     generate_kwargs = dict(
         model_inputs,
@@ -32,37 +35,42 @@ def run_generation(user_text, top_p, temperature, top_k, max_new_tokens):
         temperature=float(temperature),
         top_k=top_k
     )
+    
+    # Create a new thread for model generation
     t = Thread(target=model.generate, kwargs=generate_kwargs)
     t.start()
 
-    # Pull the generated text from the streamer, and update the model output.
     model_output = ""
     for new_text in streamer:
         model_output += new_text
         yield model_output
     return model_output
 
-
+# Define a function to reset the user input textbox
 def reset_textbox():
     return gr.update(value='')
 
-
+# Create a Gradio UI interface
 with gr.Blocks() as demo:
-    # duplicate_link = "http://huggingface.co/spaces/zubairsamo/transformers_streaming?duplicate=true"
+    # Display a title
     gr.Markdown(
         "# Testing ALPACA Model \n"
     )
 
     with gr.Row():
         with gr.Column(scale=4):
+            # Create a textbox for user input
             user_text = gr.Textbox(
                 placeholder="Ask Me Anything ... ",
                 label="User input"
             )
+            # Create a textbox for model output
             model_output = gr.Textbox(label="Model output", lines=10, interactive=False)
+            # Create a submit button
             button_submit = gr.Button(value="Submit")
 
         with gr.Column(scale=1):
+            # Create sliders for adjusting generation parameters
             max_new_tokens = gr.Slider(
                 minimum=1, maximum=1000, value=250, step=1, interactive=True, label="Max New Tokens",
             )
@@ -76,7 +84,9 @@ with gr.Blocks() as demo:
                 minimum=0.1, maximum=5.0, value=0.8, step=0.1, interactive=True, label="Temperature",
             )
 
+    # Set up the submission of user input
     user_text.submit(run_generation, [user_text, top_p, temperature, top_k, max_new_tokens], model_output)
     button_submit.click(run_generation, [user_text, top_p, temperature, top_k, max_new_tokens], model_output)
 
-    demo.queue(max_size=32).launch(enable_queue=True)
+    # Launch the Gradio interface
+    demo.queue(max_size=32).launch(enable_queue=True)