Spaces:
Runtime error
Runtime error
Add application file
Browse files
app.py
CHANGED
@@ -14,6 +14,7 @@ import re
|
|
14 |
import torch
|
15 |
import transformers
|
16 |
from transformers import pipeline
|
|
|
17 |
from datasets import load_dataset
|
18 |
import soundfile as sf
|
19 |
from IPython.display import Audio
|
@@ -144,9 +145,15 @@ def main_function(uploaded_filepath):
|
|
144 |
text_per_pagy[key] = cleaned_text
|
145 |
abstract_text = extract_abstract(text_per_pagy)
|
146 |
|
|
|
|
|
147 |
#abstract the summary with my pipeline and model, deciding the length
|
148 |
summarizer = pipeline("summarization", model="pszemraj/long-t5-tglobal-base-sci-simplify")
|
149 |
-
summary = summarizer(abstract_text, max_length=
|
|
|
|
|
|
|
|
|
150 |
|
151 |
#generating the audio from the text, with my pipeline and model
|
152 |
synthesiser = pipeline("text-to-speech", model="microsoft/speecht5_tts")
|
@@ -159,7 +166,7 @@ def main_function(uploaded_filepath):
|
|
159 |
sf.write(audio_file_path, speech["audio"], samplerate=speech["sampling_rate"])
|
160 |
|
161 |
#the function returns the 2 pieces we need
|
162 |
-
return
|
163 |
|
164 |
#let's communicate with gradio what it has to put in
|
165 |
iface = gr.Interface(
|
|
|
14 |
import torch
|
15 |
import transformers
|
16 |
from transformers import pipeline
|
17 |
+
import nltk
|
18 |
from datasets import load_dataset
|
19 |
import soundfile as sf
|
20 |
from IPython.display import Audio
|
|
|
145 |
text_per_pagy[key] = cleaned_text
|
146 |
abstract_text = extract_abstract(text_per_pagy)
|
147 |
|
148 |
+
nltk.download('punkt')
|
149 |
+
|
150 |
#abstract the summary with my pipeline and model, deciding the length
|
151 |
summarizer = pipeline("summarization", model="pszemraj/long-t5-tglobal-base-sci-simplify")
|
152 |
+
summary = summarizer(abstract_text, max_length=100, do_sample=False)[0]['summary_text']
|
153 |
+
|
154 |
+
#keeping just the first sentence, to be sure.
|
155 |
+
sentences = nltk.tokenize.sent_tokenize(summary)
|
156 |
+
first_sentence = sentences[0]
|
157 |
|
158 |
#generating the audio from the text, with my pipeline and model
|
159 |
synthesiser = pipeline("text-to-speech", model="microsoft/speecht5_tts")
|
|
|
166 |
sf.write(audio_file_path, speech["audio"], samplerate=speech["sampling_rate"])
|
167 |
|
168 |
#the function returns the 2 pieces we need
|
169 |
+
return first_sentence, audio_file_path
|
170 |
|
171 |
#let's communicate with gradio what it has to put in
|
172 |
iface = gr.Interface(
|