app.py

import gradio as gr
from transformers import pipeline
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
import torch
import warnings
import weave
import os
import wandb

wandb.login(key=os.getenv('WANDB_KEY'))
weave.init("Faseeh")

warnings.filterwarnings("ignore")
device = "cuda" if torch.cuda.is_available() else "cpu"

# Load the tokenizer and model
tokenizer = AutoTokenizer.from_pretrained("Abdulmohsena/Faseeh")
model = AutoModelForSeq2SeqLM.from_pretrained("Abdulmohsena/Faseeh").to(device)

class Faseeh(weave.Model):

    @weave.op()
    def translate(self, input_data: str, temperature: float = 0.1):
        inputs = tokenizer(input_data, return_tensors='pt')
        outputs = model.generate(**inputs,
                                 temperature=temperature,
                                 do_sample=True)[0]
        prediction = tokenizer.decode(outputs, skip_special_tokens=True)
            
        return prediction

faseeh = Faseeh()

# Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("# Faseeh: The Elequent Arabic Translator")
    
    # Input text box
    input_text = gr.Textbox(label="Input Text", placeholder="Enter text to translate from English to Classical Arabic")

    # Output text box
    output_text = gr.Textbox(label="ادخل نصا لتترجمه من الأنجليزية الى العربية الفصيحة")

    # Button to trigger translation
    translate_btn = gr.Button("Translate")

    # Button action
    translate_btn.click(faseeh.translate, inputs=input_text, outputs=output_text)

# Launch the Gradio app
if __name__ == "__main__":
    demo.launch()