Create app.py

app.py

@@ -0,0 +1,63 @@
+import os
+from datasets import load_dataset
+import pandas as pd
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+import gradio as gr
+
+# Step 1: Load dataset and save to CSV (auto-training data)
+def load_and_save_dataset():
+    print("Loading dataset from Hugging Face...")
+    dataset = load_dataset("HuggingFaceFW/fineweb", split="train")
+    print("Saving dataset to data.csv...")
+    dataset.to_csv("data.csv")
+    print("Done! Data saved to data.csv.")
+    return "Dataset loaded and saved to data.csv."
+
+# Run on startup
+load_and_save_dataset()
+
+# Step 2: Load GPT-2 model for inference
+model_name = "gpt2"  # or "distilgpt2" for faster inference
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+# Create generator pipeline
+generator = pipeline('text-generation', model=model, tokenizer=tokenizer, device=-1)
+
+# Function to generate responses
+def generate_response(prompt):
+    responses = generator(
+        prompt,
+        max_length=100,
+        do_sample=True,
+        temperature=0.7,
+        top_k=50,
+        top_p=0.95,
+        num_return_sequences=1
+    )
+    return responses[0]['generated_text'].strip()
+
+# Step 3: Build Gradio interface
+with gr.Blocks() as demo:
+    gr.Markdown("## GPT-2 Based AI Assistant with Dataset Loaded from Hugging Face")
+    gr.Textbox(value="Loading dataset...", interactive=False, lines=2)
+    fetch_button = gr.Button("Load Dataset and Save CSV")
+    output_message = gr.Textbox()
+
+    def fetch_and_confirm():
+        msg = load_and_save_dataset()
+        return msg
+
+    fetch_button.click(fetch_and_confirm, outputs=output_message)
+
+    gr.Markdown("### Ask the AI Assistant")
+    prompt_input = gr.Textbox(label="Enter your prompt", placeholder="Say something...")
+    response_output = gr.Textbox(label="Response", lines=10)
+
+    def respond(prompt):
+        return generate_response(prompt)
+
+    gr.Button("Ask").click(respond, inputs=prompt_input, outputs=response_output)
+
+demo.launch()