app.py

from transformers import GPT2LMHeadModel, GPT2Tokenizer
import gradio as gr

model_name = "gpt2"
model = GPT2LMHeadModel.from_pretrained(model_name)
tokenizer = GPT2Tokenizer.from_pretrained(model_name)

def generate(text):
    token_ids = tokenizer.encode(text, return_tensors="pt")
    gpt2_tensors = model.generate(token_ids, max_length=200, 
                                  no_repeat_ngram_size=True, 
                                  num_beams=3,
                                  do_sample=True,
                                  temperature=1.5
                                 )
    
    #response= gpt2_tensors
    response = ""
    for i, x in enumerate(gpt2_tensors):
        response += f"{i}: {tokenizer.decode(x, skip_special_tokens=True)}"
    return response

in_text = gr.Textbox(lines=1, label="English", placeholder="English text here")
out = gr.Textbox(lines=1, label="Generated tensors")
gr.Interface(generate, inputs=in_text, outputs=out).launch()