app.py

import gradio as gr
import torch
import time
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
from threading import Thread
import time
import pytz
from datetime import datetime
import gradio as gr
import torch
import time
import pytz
from datetime import datetime
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
from threading import Thread
print("Loading model and tokenizer...")
model_name = "large-traversaal/Phi-4-Hindi"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
print("Model and tokenizer loaded successfully!")
option_mapping = {
    "translation": "### TRANSLATION ###",
    "mcq": "### MCQ ###",
    "nli": "### NLI ###",
    "summarization": "### SUMMARIZATION ###",
    "long response": "### LONG RESPONSE ###",
    "short response": "### SHORT RESPONSE ###",
    "direct response": "### DIRECT RESPONSE ###",
    "paraphrase": "### PARAPHRASE ###",
    "code": "### CODE ###"
}
def generate_response(message, temperature, max_new_tokens, top_p, task):
    append_text = option_mapping.get(task, "")
    prompt = f"INPUT : {message} {append_text} RESPONSE : "
    print(f"Prompt: {prompt}")
    start_time = time.time()
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
    streamer = TextIteratorStreamer(tokenizer, skip_special_tokens=True)
    gen_kwargs = {
        "input_ids": inputs["input_ids"],
        "streamer": streamer,
        "temperature": temperature,
        "max_new_tokens": max_new_tokens,
        "top_p": top_p,
        "do_sample": True if temperature > 0 else False,
    }
    thread = Thread(target=model.generate, kwargs=gen_kwargs)
    thread.start()
    result = []
    for text in streamer:
        result.append(text)
        yield "".join(result)
    end_time = time.time()
    time_taken = end_time - start_time
    output_text = "".join(result)
    if "RESPONSE : " in output_text:
        output_text = output_text.split("RESPONSE : ", 1)[1].strip()
    print(f"Output: {output_text}")
    print(f"Time taken: {time_taken:.2f} seconds")
    pst_timezone = pytz.timezone('America/Los_Angeles')
    current_time_pst = datetime.now(pst_timezone).strftime("%Y-%m-%d %H:%M:%S %Z%z")
    print(f"Current timestamp (PST): {current_time_pst}")
with gr.Blocks() as demo:
    gr.Markdown("# Phi-4-Hindi Demo")
    with gr.Row():
        with gr.Column():
            input_text = gr.Textbox(
                label="Input",
                placeholder="Enter your text here...",
                lines=5
            )
            task_dropdown = gr.Dropdown(
                choices=["translation", "mcq", "nli", "summarization", "long response", "short response", "direct response", "paraphrase", "code"],
                value="long response",
                label="Task"
            )
            with gr.Row():
                with gr.Column():
                    temperature = gr.Slider(
                        minimum=0.0,
                        maximum=1.0,
                        value=0.1,
                        step=0.01,
                        label="Temperature"
                    )
                with gr.Column():
                    max_new_tokens = gr.Slider(
                        minimum=50,
                        maximum=1000,
                        value=400,
                        step=10,
                        label="Max New Tokens"
                    )
                with gr.Column():
                    top_p = gr.Slider(
                        minimum=0.0,
                        maximum=1.0,
                        value=0.1,
                        step=0.01,
                        label="Top P"
                    )
            with gr.Row():
                clear_btn = gr.Button("Clear")
                send_btn = gr.Button("Send", variant="primary")
        with gr.Column():
            output_text = gr.Textbox(
                label="Output",
                lines=15
            )
    send_btn.click(
        fn=generate_response,
        inputs=[input_text, temperature, max_new_tokens, top_p, task_dropdown],
        outputs=output_text
    )
    clear_btn.click(
        fn=lambda: ("", ""),
        inputs=None,
        outputs=[input_text, output_text]
    )
if __name__ == "__main__":
    demo.queue().launch()