Spaces:
Sleeping
Sleeping
Edward Baker
commited on
Commit
·
07cb201
1
Parent(s):
d6b5ce1
updating with proper gradio
Browse files
app.py
CHANGED
|
@@ -1,7 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import gradio as gr
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
|
| 3 |
-
|
| 4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
|
| 6 |
-
iface
|
| 7 |
-
iface.launch()
|
|
|
|
| 1 |
+
from transformers import LEDTokenizer, LEDForConditionalGeneration
|
| 2 |
+
import torch
|
| 3 |
+
import re
|
| 4 |
+
tokenizer = LEDTokenizer.from_pretrained("patrickvonplaten/led-large-16384-pubmed")
|
| 5 |
+
model = LEDForConditionalGeneration.from_pretrained("patrickvonplaten/led-large-16384-pubmed").to("cuda").half()
|
| 6 |
+
|
| 7 |
import gradio as gr
|
| 8 |
+
import os
|
| 9 |
+
import docx2txt
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
tokenizer = LEDTokenizer.from_pretrained("patrickvonplaten/led-large-16384-pubmed")
|
| 13 |
+
model = LEDForConditionalGeneration.from_pretrained("patrickvonplaten/led-large-16384-pubmed", return_dict_in_generate=True).to("cuda")
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
def summarize(text_file):
|
| 18 |
+
file_extension = os.path.splitext(text_file.name)[1]
|
| 19 |
+
if file_extension == ".txt":
|
| 20 |
+
# Load text from a txt file
|
| 21 |
+
with open(text_file.name, "r", encoding="utf-8") as f:
|
| 22 |
+
text = f.read()
|
| 23 |
+
elif file_extension == ".docx":
|
| 24 |
+
# Load text from a Word file
|
| 25 |
+
text = docx2txt.process(text_file.name)
|
| 26 |
+
else:
|
| 27 |
+
raise ValueError(f"Unsupported file type: {file_extension}")
|
| 28 |
+
|
| 29 |
+
input_ids = tokenizer(text, return_tensors="pt").input_ids.to("cuda")
|
| 30 |
+
global_attention_mask = torch.zeros_like(input_ids)
|
| 31 |
+
# set global_attention_mask on first token
|
| 32 |
+
global_attention_mask[:, 0] = 1
|
| 33 |
+
|
| 34 |
+
sequences = model.generate(input_ids, global_attention_mask=global_attention_mask).sequences
|
| 35 |
+
|
| 36 |
+
summary = tokenizer.batch_decode(sequences)[0]
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
return text, summary
|
| 41 |
+
|
| 42 |
+
|
| 43 |
|
| 44 |
+
iface = gr.Interface(
|
| 45 |
+
fn=summarize,
|
| 46 |
+
inputs=gr.inputs.File(label="Upload a txt file or a Word file for the input text"),
|
| 47 |
+
outputs=[gr.outputs.Textbox(label="Original text"), gr.outputs.Textbox(label="Summary")],
|
| 48 |
+
title="Academic Paper Summarization Demo",
|
| 49 |
+
description="Upload a txt file or a Word file for the input text. Get a summary generated by a small T5 model from Hugging Face.",
|
| 50 |
+
)
|
| 51 |
|
| 52 |
+
iface.launch()
|
|
|