mmcquade's picture
updates to app.py
4629047
raw
history blame
725 Bytes
import gradio as gr
import torch
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("mmcquade11/autonlp-reuters-summarization-34018133")
model = AutoModelForSeq2SeqLM.from_pretrained("mmcquade11/autonlp-reuters-summarization-34018133")
def summarize(text):
input_ids = torch.tensor(tokenizer.encode(text, add_special_tokens=True)).unsqueeze(0)
summary_ids = model.generate(input_ids, num_beams=4, max_length=100, early_stopping=True)
return tokenizer.decode(summary_ids[0], skip_special_tokens=True)
def summarize_text(text):
return summarize(text)
iface = gr.Interface(summarize_text, "textbox", "label")
if __name__ == "__main__":
iface.launch()