updates to files

app.py

@@ -1,37 +1,18 @@
 #python3
+#pytorch
 #build a text summarizer using hugging face and gradio
 
-
 import gradio as gr
-import pandas as pd
-import numpy as np
-import tensorflow as tf
 import transformers
-from transformers import TFAutoModel, AutoTokenizer
-
-model_class, tokenizer_class, pretrained_weights = (TFAutoModel, AutoTokenizer, 'bert-base-uncased')
-
-# Load pretrained model/tokenizer
-tokenizer = tokenizer_class.from_pretrained(pretrained_weights)
-model = model_class.from_pretrained(pretrained_weights)
-
-def get_summary(article):
-    article_input_ids = tokenizer.encode(article, return_tensors='tf')
-    summary_ids = model.generate(article_input_ids)
-    summary_txt = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
-    return summary_txt
+from transformers import BartTokenizer, BartForConditionalGeneration
 
-def get_summary_gradio(article):
-    return get_summary(article)
+model = BartForConditionalGeneration.from_pretrained('facebook/bart-large-cnn')
+tokenizer = BartTokenizer.from_pretrained('facebook/bart-large-cnn')
 
-iface = gr.Interface(get_summary_gradio, "textbox", "textbox", live=True,
-                     examples=[
-                         ["The quick brown fox jumps over the lazy dog."],
-                         ["The world is a strange place. Sometimes, things are what they seem. But then, if you look closer, they can become something entirely different."],
-                         ["The sky is clear; the stars are twinkling. I'm going to bed now. Good night."],
-                         ["The president of the United States, and the president of the United Kingdom, have both been in the White House."],
-                         ["The president of the United States, and the president of the United Kingdom, have both been in the White House."]
-                     ])
+def bart_summarizer(input_text):
+	input_text = tokenizer.batch_encode_plus([input_text], max_length=1024, return_tensors='pt')
+	summary_ids = model.generate(input_text['input_ids'], num_beams=4, max_length=100, early_stopping=True)
+	output = [tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=False) for g in summary_ids]
+	return output[0]
 
-if __name__ == "__main__":
-    iface.launch()
+gr.Interface(fn=bart_summarizer, inputs=gr.inputs.Textbox(lines=7, placeholder="Enter some long text here"), outputs="textbox", live=True).launch()

requirements.txt

@@ -1,5 +1,2 @@
-tensorflow
-gradio
-numpy
 transformers
 torch