import gradio as gr
import tensorflow as tf
from transformers import TFGPT2LMHeadModel, GPT2Tokenizer

tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
model = TFGPT2LMHeadModel.from_pretrained("gpt2", pad_token_id=tokenizer.eos_token_id)

#uncomment the lines below if you want to use GPT2-XL instead of GPT2 which would give faster outputs with greater accuracy.
#you will have to recomment the above 2 lines if you wanna use GPT2-XL.
#tokenizer = GPT2Tokenizer.from_pretrained("gpt2-xl")
#model = TFGPT2LMHeadModel.from_pretrained("gpt2-xl", pad_token_id=tokenizer.eos_token_id)

def generate_text(inp):
    input_ids = tokenizer.encode(inp, return_tensors='tf')
    beam_output = model.generate(input_ids, max_length=100, num_beams=5, no_repeat_ngram_size=2, early_stopping=True)
    output = tokenizer.decode(beam_output[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)
    return ".".join(output.split(".")[:-1]) + "."

output_text = gr.outputs.Textbox()
gr.Interface(generate_text,"textbox", output_text, title="GPT-2",
             description="OpenAI's GPT-2 is an unsupervised language model that \
             can generate coherent text. Go ahead and input a sentence and see what it completes \
             it with! Takes around 30s to run.").launch()