#import gradio as gr
#from transformers import pipeline
#from fairseq.models.transformer import TransformerModel


# Load the English to Urdu translation model from the transformers library
#model_name_or_path = "Helsinki-NLP/opus-mt-en-ur"
#model_name_or_path = TransformerModel.from_pretrained('samiulhaq/iwslt-bt-en-ur')

#translator = pipeline("translation", model=model_name_or_path, tokenizer=model_name_or_path)

# Create a Gradio interface for the translation app
#def translate(text):
    # Use the translator pipeline to translate the input text
   # result = translator(text, max_length=500)
   # return result[0]['translation_text']

#input_text = gr.inputs.Textbox(label="Input English Text")
#output_text = gr.outputs.Textbox(label="Output Urdu Text")
#app = gr.Interface(fn=translate, inputs=input_text, outputs=output_text)

# Launch the app
#app.launch()


import gradio as gr
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

# Load the English to Urdu translation model from the transformers library
model_name_or_path = "aryanc55/english-urdu"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name_or_path)

# Create a Gradio interface for the translation app
def translate(text):
    # Tokenize the input text
    inputs = tokenizer(text, return_tensors="pt")

    # Use the model to generate the translated text
    outputs = model.generate(inputs["input_ids"], max_length=500, early_stopping=True)
    translated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)

    return translated_text

input_text = gr.inputs.Textbox(label="Input English Text")
output_text = gr.outputs.Textbox(label="Output Urdu Text")
app = gr.Interface(fn=translate, inputs=input_text, outputs=output_text)

# Launch the app
app.launch()