Spaces:
Running
on
Zero
Running
on
Zero
Implemented Chunking
Browse files
app.py
CHANGED
@@ -3,6 +3,7 @@ import pickle
|
|
3 |
import whisper
|
4 |
import streamlit as st
|
5 |
import torchaudio as ta
|
|
|
6 |
|
7 |
from io import BytesIO
|
8 |
from transformers import WhisperProcessor, WhisperForConditionalGeneration
|
@@ -12,10 +13,11 @@ device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
|
12 |
torch_dtype = torch.float16 if device == "cuda:0" else torch.float32
|
13 |
|
14 |
SAMPLING_RATE = 16000
|
|
|
15 |
|
16 |
# Load Whisper model and processor
|
17 |
processor = WhisperProcessor.from_pretrained("openai/whisper-small")
|
18 |
-
model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small")
|
19 |
|
20 |
# Title of the app
|
21 |
st.title("Audio Player with Live Transcription")
|
@@ -36,18 +38,42 @@ if 'audio_files' not in st.session_state:
|
|
36 |
|
37 |
def detect_language(audio_file):
|
38 |
whisper_model = whisper.load_model("small")
|
39 |
-
trimmed_audio = whisper.pad_or_trim(audio_file)
|
40 |
mel = whisper.log_mel_spectrogram(trimmed_audio).to(whisper_model.device)
|
41 |
-
_, probs = whisper_model.detect_language(mel
|
42 |
-
detected_lang = max(probs, key=probs.get)
|
43 |
print(f"Detected language: {detected_lang}")
|
44 |
return detected_lang
|
45 |
|
46 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
# Process uploaded files
|
48 |
if submit_button and uploaded_files is not None:
|
49 |
st.session_state.audio_files = uploaded_files
|
50 |
st.session_state.detected_languages = []
|
|
|
51 |
|
52 |
for uploaded_file in uploaded_files:
|
53 |
waveform, sampling_rate = ta.load(BytesIO(uploaded_file.read()))
|
@@ -69,30 +95,25 @@ if 'audio_files' in st.session_state and st.session_state.audio_files:
|
|
69 |
st.write(f"**Detected Language**: {st.session_state.detected_languages[i]}")
|
70 |
|
71 |
with col2:
|
72 |
-
# import pdb;pdb.set_trace()
|
73 |
-
input_features = processor(st.session_state.waveforms[i][0], sampling_rate=SAMPLING_RATE, return_tensors='pt').input_features
|
74 |
-
|
75 |
if st.button(f"Transcribe {uploaded_file.name}"):
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
|
80 |
if st.session_state.transcriptions.get(i):
|
81 |
st.write("**Transcription**:")
|
82 |
-
|
83 |
-
st.write(line)
|
84 |
|
85 |
if st.button(f"Translate {uploaded_file.name}"):
|
86 |
-
with
|
87 |
-
|
88 |
-
|
|
|
89 |
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
st.session_state.translations[i] = translation
|
94 |
|
95 |
if st.session_state.translations.get(i):
|
96 |
st.write("**Translation**:")
|
97 |
-
|
98 |
-
st.write(line)
|
|
|
3 |
import whisper
|
4 |
import streamlit as st
|
5 |
import torchaudio as ta
|
6 |
+
import numpy as np
|
7 |
|
8 |
from io import BytesIO
|
9 |
from transformers import WhisperProcessor, WhisperForConditionalGeneration
|
|
|
13 |
torch_dtype = torch.float16 if device == "cuda:0" else torch.float32
|
14 |
|
15 |
SAMPLING_RATE = 16000
|
16 |
+
CHUNK_LENGTH_S = 20 # 30 seconds per chunk
|
17 |
|
18 |
# Load Whisper model and processor
|
19 |
processor = WhisperProcessor.from_pretrained("openai/whisper-small")
|
20 |
+
model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small").to(device)
|
21 |
|
22 |
# Title of the app
|
23 |
st.title("Audio Player with Live Transcription")
|
|
|
38 |
|
39 |
def detect_language(audio_file):
|
40 |
whisper_model = whisper.load_model("small")
|
41 |
+
trimmed_audio = whisper.pad_or_trim(audio_file.squeeze())
|
42 |
mel = whisper.log_mel_spectrogram(trimmed_audio).to(whisper_model.device)
|
43 |
+
_, probs = whisper_model.detect_language(mel)
|
44 |
+
detected_lang = max(probs[0], key=probs[0].get)
|
45 |
print(f"Detected language: {detected_lang}")
|
46 |
return detected_lang
|
47 |
|
48 |
|
49 |
+
def process_long_audio(waveform, sampling_rate, task="transcribe", language=None):
|
50 |
+
input_length = waveform.shape[1]
|
51 |
+
chunk_length = int(CHUNK_LENGTH_S * sampling_rate)
|
52 |
+
chunks = [waveform[:, i:i + chunk_length] for i in range(0, input_length, chunk_length)]
|
53 |
+
|
54 |
+
results = []
|
55 |
+
for chunk in chunks:
|
56 |
+
# import pdb;pdb.set_trace()
|
57 |
+
input_features = processor(chunk[0], sampling_rate=sampling_rate, return_tensors="pt").input_features.to(device)
|
58 |
+
|
59 |
+
with torch.no_grad():
|
60 |
+
if task == "translate":
|
61 |
+
forced_decoder_ids = processor.get_decoder_prompt_ids(language=language, task="translate")
|
62 |
+
generated_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids)
|
63 |
+
else:
|
64 |
+
generated_ids = model.generate(input_features)
|
65 |
+
|
66 |
+
transcription = processor.batch_decode(generated_ids, skip_special_tokens=True)
|
67 |
+
results.extend(transcription)
|
68 |
+
|
69 |
+
return " ".join(results)
|
70 |
+
|
71 |
+
|
72 |
# Process uploaded files
|
73 |
if submit_button and uploaded_files is not None:
|
74 |
st.session_state.audio_files = uploaded_files
|
75 |
st.session_state.detected_languages = []
|
76 |
+
st.session_state.waveforms = []
|
77 |
|
78 |
for uploaded_file in uploaded_files:
|
79 |
waveform, sampling_rate = ta.load(BytesIO(uploaded_file.read()))
|
|
|
95 |
st.write(f"**Detected Language**: {st.session_state.detected_languages[i]}")
|
96 |
|
97 |
with col2:
|
|
|
|
|
|
|
98 |
if st.button(f"Transcribe {uploaded_file.name}"):
|
99 |
+
with st.spinner("Transcribing..."):
|
100 |
+
transcription = process_long_audio(st.session_state.waveforms[i], SAMPLING_RATE)
|
101 |
+
st.session_state.transcriptions[i] = transcription
|
102 |
|
103 |
if st.session_state.transcriptions.get(i):
|
104 |
st.write("**Transcription**:")
|
105 |
+
st.write(st.session_state.transcriptions[i])
|
|
|
106 |
|
107 |
if st.button(f"Translate {uploaded_file.name}"):
|
108 |
+
with st.spinner("Translating..."):
|
109 |
+
with open('languages.pkl', 'rb') as f:
|
110 |
+
lang_dict = pickle.load(f)
|
111 |
+
detected_language_name = lang_dict[st.session_state.detected_languages[i]]
|
112 |
|
113 |
+
translation = process_long_audio(st.session_state.waveforms[i], SAMPLING_RATE, task="translate",
|
114 |
+
language=detected_language_name)
|
115 |
+
st.session_state.translations[i] = translation
|
|
|
116 |
|
117 |
if st.session_state.translations.get(i):
|
118 |
st.write("**Translation**:")
|
119 |
+
st.write(st.session_state.translations[i])
|
|