Spaces:
Runtime error
Runtime error
suisuyy
commited on
Commit
·
a1dd53c
1
Parent(s):
fbe6007
add status
Browse files- README.md +293 -0
- __pycache__/app.cpython-310.pyc +0 -0
- app.py +79 -7
- requirements.txt +4 -1
README.md
CHANGED
|
@@ -10,3 +10,296 @@ pinned: false
|
|
| 10 |
---
|
| 11 |
|
| 12 |
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
# all requirements when dev
|
| 16 |
+
```
|
| 17 |
+
|
| 18 |
+
absl-py==2.0.0
|
| 19 |
+
aiofiles==22.1.0
|
| 20 |
+
aiohttp==3.9.1
|
| 21 |
+
aiosignal==1.3.1
|
| 22 |
+
aiosqlite==0.19.0
|
| 23 |
+
altair==5.3.0
|
| 24 |
+
annotated-types==0.6.0
|
| 25 |
+
anyascii==0.3.2
|
| 26 |
+
anyio==4.2.0
|
| 27 |
+
argon2-cffi==23.1.0
|
| 28 |
+
argon2-cffi-bindings==21.2.0
|
| 29 |
+
arrow==1.3.0
|
| 30 |
+
asttokens==2.4.1
|
| 31 |
+
async-timeout==4.0.3
|
| 32 |
+
attrs==23.2.0
|
| 33 |
+
audioread==3.0.1
|
| 34 |
+
Babel==2.14.0
|
| 35 |
+
backoff==2.2.1
|
| 36 |
+
bangla==0.0.2
|
| 37 |
+
beautifulsoup4==4.12.2
|
| 38 |
+
bleach==6.1.0
|
| 39 |
+
blessed==1.20.0
|
| 40 |
+
blinker==1.7.0
|
| 41 |
+
blis==0.7.11
|
| 42 |
+
bnnumerizer==0.0.2
|
| 43 |
+
bnunicodenormalizer==0.1.6
|
| 44 |
+
boto3==1.34.17
|
| 45 |
+
botocore==1.34.17
|
| 46 |
+
Brotli==1.1.0
|
| 47 |
+
cachetools==5.3.2
|
| 48 |
+
catalogue==2.0.10
|
| 49 |
+
certifi==2023.11.17
|
| 50 |
+
cffi==1.16.0
|
| 51 |
+
charset-normalizer==3.3.2
|
| 52 |
+
click==8.1.7
|
| 53 |
+
cloudpathlib==0.16.0
|
| 54 |
+
colorama==0.4.6
|
| 55 |
+
comm==0.2.1
|
| 56 |
+
confection==0.1.4
|
| 57 |
+
contourpy==1.2.0
|
| 58 |
+
coqpit==0.0.17
|
| 59 |
+
cutlet==0.3.0
|
| 60 |
+
cycler==0.12.1
|
| 61 |
+
cymem==2.0.8
|
| 62 |
+
Cython==3.0.8
|
| 63 |
+
dateparser==1.1.8
|
| 64 |
+
debugpy==1.8.0
|
| 65 |
+
decorator==4.4.2
|
| 66 |
+
deepspeed==0.12.6
|
| 67 |
+
defusedxml==0.7.1
|
| 68 |
+
docopt==0.6.2
|
| 69 |
+
einops==0.7.0
|
| 70 |
+
emoji==2.8.0
|
| 71 |
+
encodec==0.1.1
|
| 72 |
+
entrypoints==0.4
|
| 73 |
+
exceptiongroup==1.2.0
|
| 74 |
+
executing==2.0.1
|
| 75 |
+
fastapi==0.109.0
|
| 76 |
+
fastjsonschema==2.19.1
|
| 77 |
+
ffmpy==0.3.2
|
| 78 |
+
filelock==3.13.1
|
| 79 |
+
Flask==3.0.0
|
| 80 |
+
fonttools==4.47.2
|
| 81 |
+
fqdn==1.5.1
|
| 82 |
+
frozenlist==1.4.1
|
| 83 |
+
fsspec==2023.12.2
|
| 84 |
+
fugashi==1.3.0
|
| 85 |
+
g2pkk==0.1.2
|
| 86 |
+
google-auth==2.26.2
|
| 87 |
+
google-auth-oauthlib==1.2.0
|
| 88 |
+
gpustat==1.1.1
|
| 89 |
+
gradio==4.8.0
|
| 90 |
+
gradio_client==0.10.0
|
| 91 |
+
grpcio==1.60.0
|
| 92 |
+
gruut==2.2.3
|
| 93 |
+
gruut-ipa==0.13.0
|
| 94 |
+
gruut-lang-de==2.0.0
|
| 95 |
+
gruut-lang-en==2.0.0
|
| 96 |
+
gruut-lang-es==2.0.0
|
| 97 |
+
gruut-lang-fr==2.0.2
|
| 98 |
+
h11==0.14.0
|
| 99 |
+
hangul-romanize==0.1.0
|
| 100 |
+
hjson==3.1.0
|
| 101 |
+
httpcore==1.0.2
|
| 102 |
+
httpx==0.26.0
|
| 103 |
+
huggingface-hub==0.20.2
|
| 104 |
+
idna==3.6
|
| 105 |
+
imageio==2.34.0
|
| 106 |
+
imageio-ffmpeg==0.4.9
|
| 107 |
+
importlib_resources==6.4.0
|
| 108 |
+
inflect==7.0.0
|
| 109 |
+
ipykernel==6.26.0
|
| 110 |
+
ipython==8.17.2
|
| 111 |
+
ipython-genutils==0.2.0
|
| 112 |
+
ipywidgets==8.1.1
|
| 113 |
+
isoduration==20.11.0
|
| 114 |
+
itsdangerous==2.1.2
|
| 115 |
+
jaconv==0.3.4
|
| 116 |
+
jamo==0.4.1
|
| 117 |
+
jedi==0.19.1
|
| 118 |
+
jieba==0.42.1
|
| 119 |
+
Jinja2==3.1.3
|
| 120 |
+
jmespath==1.0.1
|
| 121 |
+
joblib==1.3.2
|
| 122 |
+
json5==0.9.14
|
| 123 |
+
jsonlines==1.2.0
|
| 124 |
+
jsonpointer==2.4
|
| 125 |
+
jsonschema==4.20.0
|
| 126 |
+
jsonschema-specifications==2023.12.1
|
| 127 |
+
jupyter-events==0.9.0
|
| 128 |
+
jupyter-ydoc==0.2.5
|
| 129 |
+
jupyter_client==7.4.9
|
| 130 |
+
jupyter_core==5.7.1
|
| 131 |
+
jupyter_server==2.12.4
|
| 132 |
+
jupyter_server_fileid==0.9.1
|
| 133 |
+
jupyter_server_terminals==0.5.1
|
| 134 |
+
jupyter_server_ydoc==0.6.1
|
| 135 |
+
jupyterlab==3.6.1
|
| 136 |
+
jupyterlab-widgets==3.0.9
|
| 137 |
+
jupyterlab_pygments==0.3.0
|
| 138 |
+
jupyterlab_server==2.25.2
|
| 139 |
+
kiwisolver==1.4.5
|
| 140 |
+
langcodes==3.3.0
|
| 141 |
+
lazy_loader==0.3
|
| 142 |
+
librosa==0.10.0
|
| 143 |
+
lightning==2.1.2
|
| 144 |
+
lightning-cloud==0.5.57
|
| 145 |
+
lightning-utilities==0.10.0
|
| 146 |
+
lightning_sdk==0.0.13a0
|
| 147 |
+
linkify-it-py==2.0.3
|
| 148 |
+
llvmlite==0.41.1
|
| 149 |
+
loguru==0.7.2
|
| 150 |
+
Markdown==3.5.2
|
| 151 |
+
markdown-it-py==2.2.0
|
| 152 |
+
MarkupSafe==2.1.3
|
| 153 |
+
matplotlib==3.8.2
|
| 154 |
+
matplotlib-inline==0.1.6
|
| 155 |
+
mdit-py-plugins==0.3.3
|
| 156 |
+
mdurl==0.1.2
|
| 157 |
+
mistune==3.0.2
|
| 158 |
+
mojimoji==0.0.13
|
| 159 |
+
moviepy==1.0.3
|
| 160 |
+
mpmath==1.3.0
|
| 161 |
+
msgpack==1.0.7
|
| 162 |
+
multidict==6.0.4
|
| 163 |
+
murmurhash==1.0.10
|
| 164 |
+
mutagen==1.47.0
|
| 165 |
+
nbclassic==1.0.0
|
| 166 |
+
nbclient==0.9.0
|
| 167 |
+
nbconvert==7.14.1
|
| 168 |
+
nbformat==5.9.2
|
| 169 |
+
nest-asyncio==1.5.8
|
| 170 |
+
networkx==2.8.8
|
| 171 |
+
ninja==1.11.1.1
|
| 172 |
+
nltk==3.8.1
|
| 173 |
+
notebook==6.5.6
|
| 174 |
+
notebook_shim==0.2.3
|
| 175 |
+
num2words==0.5.13
|
| 176 |
+
numba==0.58.1
|
| 177 |
+
numpy==1.22.0
|
| 178 |
+
nvidia-ml-py==12.535.133
|
| 179 |
+
oauthlib==3.2.2
|
| 180 |
+
orjson==3.10.0
|
| 181 |
+
overrides==7.4.0
|
| 182 |
+
packaging==23.2
|
| 183 |
+
pandas==1.5.3
|
| 184 |
+
pandocfilters==1.5.0
|
| 185 |
+
parso==0.8.3
|
| 186 |
+
pexpect==4.9.0
|
| 187 |
+
pillow==10.2.0
|
| 188 |
+
platformdirs==4.1.0
|
| 189 |
+
pooch==1.8.0
|
| 190 |
+
preshed==3.0.9
|
| 191 |
+
proglog==0.1.10
|
| 192 |
+
prometheus-client==0.19.0
|
| 193 |
+
prompt-toolkit==3.0.43
|
| 194 |
+
protobuf==4.23.4
|
| 195 |
+
psutil==5.9.7
|
| 196 |
+
ptyprocess==0.7.0
|
| 197 |
+
pure-eval==0.2.2
|
| 198 |
+
py-cpuinfo==9.0.0
|
| 199 |
+
pyasn1==0.5.1
|
| 200 |
+
pyasn1-modules==0.3.0
|
| 201 |
+
PyAudio==0.2.14
|
| 202 |
+
pycparser==2.21
|
| 203 |
+
pycryptodomex==3.20.0
|
| 204 |
+
pydantic==2.5.3
|
| 205 |
+
pydantic_core==2.14.6
|
| 206 |
+
pydub==0.25.1
|
| 207 |
+
Pygments==2.17.2
|
| 208 |
+
PyJWT==2.8.0
|
| 209 |
+
pynndescent==0.5.11
|
| 210 |
+
pynvml==11.5.0
|
| 211 |
+
pyparsing==3.1.1
|
| 212 |
+
pypinyin==0.50.0
|
| 213 |
+
pysbd==0.3.4
|
| 214 |
+
python-crfsuite==0.9.10
|
| 215 |
+
python-dateutil==2.8.2
|
| 216 |
+
python-dotenv==1.0.0
|
| 217 |
+
python-json-logger==2.0.7
|
| 218 |
+
python-multipart==0.0.6
|
| 219 |
+
pytorch-lightning==2.1.2
|
| 220 |
+
pyttsx3==2.90
|
| 221 |
+
pytz==2023.3.post1
|
| 222 |
+
PyYAML==6.0.1
|
| 223 |
+
pyzmq==24.0.1
|
| 224 |
+
referencing==0.32.1
|
| 225 |
+
regex==2023.12.25
|
| 226 |
+
requests==2.31.0
|
| 227 |
+
requests-oauthlib==1.3.1
|
| 228 |
+
rfc3339-validator==0.1.4
|
| 229 |
+
rfc3986-validator==0.1.1
|
| 230 |
+
rich==13.7.0
|
| 231 |
+
rpds-py==0.17.1
|
| 232 |
+
rsa==4.9
|
| 233 |
+
s3transfer==0.10.0
|
| 234 |
+
safetensors==0.4.1
|
| 235 |
+
scikit-learn==1.3.2
|
| 236 |
+
scipy==1.11.4
|
| 237 |
+
semantic-version==2.10.0
|
| 238 |
+
Send2Trash==1.8.2
|
| 239 |
+
shellingham==1.5.4
|
| 240 |
+
six==1.16.0
|
| 241 |
+
smart-open==6.4.0
|
| 242 |
+
sniffio==1.3.0
|
| 243 |
+
soundfile==0.12.1
|
| 244 |
+
soupsieve==2.5
|
| 245 |
+
soxr==0.3.7
|
| 246 |
+
spaces==0.25.0
|
| 247 |
+
spacy==3.7.2
|
| 248 |
+
spacy-legacy==3.0.12
|
| 249 |
+
spacy-loggers==1.0.5
|
| 250 |
+
srsly==2.4.8
|
| 251 |
+
stack-data==0.6.3
|
| 252 |
+
stanza==1.6.1
|
| 253 |
+
starlette==0.35.1
|
| 254 |
+
stream2sentence==0.2.2
|
| 255 |
+
SudachiDict-core==20230927
|
| 256 |
+
SudachiPy==0.6.8
|
| 257 |
+
sympy==1.12
|
| 258 |
+
tensorboard==2.15.1
|
| 259 |
+
tensorboard-data-server==0.7.2
|
| 260 |
+
terminado==0.18.0
|
| 261 |
+
thinc==8.2.2
|
| 262 |
+
threadpoolctl==3.2.0
|
| 263 |
+
tinycss2==1.2.1
|
| 264 |
+
tokenizers==0.15.0
|
| 265 |
+
tomli==2.0.1
|
| 266 |
+
tomlkit==0.12.0
|
| 267 |
+
toolz==0.12.1
|
| 268 |
+
torch==2.1.1+cu118
|
| 269 |
+
torchaudio==2.1.1+cu118
|
| 270 |
+
torchmetrics==1.2.0
|
| 271 |
+
torchvision==0.16.1+cu121
|
| 272 |
+
tornado==6.4
|
| 273 |
+
tqdm==4.66.1
|
| 274 |
+
trainer==0.0.36
|
| 275 |
+
traitlets==5.14.1
|
| 276 |
+
transformers @ git+https://github.com/huggingface/transformers@81642d2b51de9d5e5aee1768abdc744d90f7f52d
|
| 277 |
+
triton==2.1.0
|
| 278 |
+
TTS==0.21.3
|
| 279 |
+
typer==0.9.0
|
| 280 |
+
types-python-dateutil==2.8.19.20240106
|
| 281 |
+
typing_extensions==4.9.0
|
| 282 |
+
tzdata==2023.4
|
| 283 |
+
tzlocal==5.2
|
| 284 |
+
uc-micro-py==1.0.3
|
| 285 |
+
umap-learn==0.5.5
|
| 286 |
+
Unidecode==1.3.8
|
| 287 |
+
unidic-lite==1.0.8
|
| 288 |
+
uri-template==1.3.0
|
| 289 |
+
urllib3==2.0.7
|
| 290 |
+
uvicorn==0.25.0
|
| 291 |
+
wasabi==1.1.2
|
| 292 |
+
wcwidth==0.2.13
|
| 293 |
+
weasel==0.3.4
|
| 294 |
+
webcolors==1.13
|
| 295 |
+
webencodings==0.5.1
|
| 296 |
+
websocket-client==1.7.0
|
| 297 |
+
websockets==11.0.3
|
| 298 |
+
Werkzeug==3.0.1
|
| 299 |
+
widgetsnbextension==4.0.9
|
| 300 |
+
xtts-api-server==0.8.3
|
| 301 |
+
y-py==0.6.2
|
| 302 |
+
yarl==1.9.4
|
| 303 |
+
ypy-websocket==0.8.4
|
| 304 |
+
yt-dlp==2024.3.10
|
| 305 |
+
```
|
__pycache__/app.cpython-310.pyc
CHANGED
|
Binary files a/__pycache__/app.cpython-310.pyc and b/__pycache__/app.cpython-310.pyc differ
|
|
|
app.py
CHANGED
|
@@ -1,16 +1,18 @@
|
|
| 1 |
import torch
|
| 2 |
import time
|
| 3 |
-
|
|
|
|
| 4 |
import gradio as gr
|
| 5 |
import spaces
|
| 6 |
from transformers import pipeline
|
| 7 |
from transformers.pipelines.audio_utils import ffmpeg_read
|
| 8 |
|
| 9 |
DEFAULT_MODEL_NAME = "distil-whisper/distil-large-v3"
|
| 10 |
-
|
| 11 |
BATCH_SIZE = 8
|
| 12 |
|
| 13 |
device = 0 if torch.cuda.is_available() else "cpu"
|
|
|
|
|
|
|
| 14 |
|
| 15 |
def load_pipeline(model_name):
|
| 16 |
return pipeline(
|
|
@@ -32,6 +34,11 @@ def transcribe(inputs, task, model_name):
|
|
| 32 |
pipe = load_pipeline(model_name)
|
| 33 |
|
| 34 |
start_time = time.time() # Record the start time
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
|
| 36 |
end_time = time.time() # Record the end time
|
| 37 |
|
|
@@ -40,12 +47,65 @@ def transcribe(inputs, task, model_name):
|
|
| 40 |
# Create the transcription time output with additional information
|
| 41 |
transcription_time_output = (
|
| 42 |
f"Transcription Time: {transcription_time:.2f} seconds\n"
|
|
|
|
| 43 |
f"Model Used: {model_name}\n"
|
| 44 |
f"Device Used: {'GPU' if torch.cuda.is_available() else 'CPU'}"
|
| 45 |
)
|
| 46 |
|
| 47 |
return text, transcription_time_output
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
demo = gr.Blocks()
|
| 50 |
|
| 51 |
mf_transcribe = gr.Interface(
|
|
@@ -57,7 +117,7 @@ mf_transcribe = gr.Interface(
|
|
| 57 |
label="Model Name",
|
| 58 |
value=DEFAULT_MODEL_NAME,
|
| 59 |
placeholder="Enter the model name",
|
| 60 |
-
info="Some available models: distil-whisper/distil-large-v3 distil-whisper/distil-medium.en
|
| 61 |
),
|
| 62 |
],
|
| 63 |
outputs=[gr.TextArea(label="Transcription"), gr.TextArea(label="Transcription Info")],
|
|
@@ -79,7 +139,7 @@ file_transcribe = gr.Interface(
|
|
| 79 |
label="Model Name",
|
| 80 |
value=DEFAULT_MODEL_NAME,
|
| 81 |
placeholder="Enter the model name",
|
| 82 |
-
info="Some available models: openai/whisper-tiny, openai/whisper-base, openai/whisper-medium, openai/whisper-large-v2"
|
| 83 |
),
|
| 84 |
],
|
| 85 |
outputs=[gr.TextArea(label="Transcription"), gr.TextArea(label="Transcription Info")],
|
|
@@ -91,8 +151,20 @@ file_transcribe = gr.Interface(
|
|
| 91 |
),
|
| 92 |
allow_flagging="never",
|
| 93 |
)
|
| 94 |
-
|
| 95 |
with demo:
|
| 96 |
gr.TabbedInterface([mf_transcribe, file_transcribe], ["Microphone", "Audio file"])
|
| 97 |
-
|
| 98 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import torch
|
| 2 |
import time
|
| 3 |
+
import moviepy.editor as mp
|
| 4 |
+
import psutil
|
| 5 |
import gradio as gr
|
| 6 |
import spaces
|
| 7 |
from transformers import pipeline
|
| 8 |
from transformers.pipelines.audio_utils import ffmpeg_read
|
| 9 |
|
| 10 |
DEFAULT_MODEL_NAME = "distil-whisper/distil-large-v3"
|
|
|
|
| 11 |
BATCH_SIZE = 8
|
| 12 |
|
| 13 |
device = 0 if torch.cuda.is_available() else "cpu"
|
| 14 |
+
if device == "cpu":
|
| 15 |
+
DEFAULT_MODEL_NAME = "openai/whisper-tiny"
|
| 16 |
|
| 17 |
def load_pipeline(model_name):
|
| 18 |
return pipeline(
|
|
|
|
| 34 |
pipe = load_pipeline(model_name)
|
| 35 |
|
| 36 |
start_time = time.time() # Record the start time
|
| 37 |
+
|
| 38 |
+
# Load the audio file and calculate its duration
|
| 39 |
+
audio = mp.AudioFileClip(inputs)
|
| 40 |
+
audio_duration = audio.duration
|
| 41 |
+
|
| 42 |
text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
|
| 43 |
end_time = time.time() # Record the end time
|
| 44 |
|
|
|
|
| 47 |
# Create the transcription time output with additional information
|
| 48 |
transcription_time_output = (
|
| 49 |
f"Transcription Time: {transcription_time:.2f} seconds\n"
|
| 50 |
+
f"Audio Duration: {audio_duration:.2f} seconds\n"
|
| 51 |
f"Model Used: {model_name}\n"
|
| 52 |
f"Device Used: {'GPU' if torch.cuda.is_available() else 'CPU'}"
|
| 53 |
)
|
| 54 |
|
| 55 |
return text, transcription_time_output
|
| 56 |
|
| 57 |
+
from gpustat import GPUStatCollection
|
| 58 |
+
|
| 59 |
+
def update_gpu_status():
|
| 60 |
+
if torch.cuda.is_available() == False:
|
| 61 |
+
return "No Nviadia Device"
|
| 62 |
+
try:
|
| 63 |
+
gpu_stats = GPUStatCollection.new_query()
|
| 64 |
+
for gpu in gpu_stats:
|
| 65 |
+
# Assuming you want to monitor the first GPU, index 0
|
| 66 |
+
gpu_id = gpu.index
|
| 67 |
+
gpu_name = gpu.name
|
| 68 |
+
gpu_utilization = gpu.utilization
|
| 69 |
+
memory_used = gpu.memory_used
|
| 70 |
+
memory_total = gpu.memory_total
|
| 71 |
+
memory_utilization = (memory_used / memory_total) * 100
|
| 72 |
+
gpu_status=(f"GPU {gpu_id}: {gpu_name}, Utilization: {gpu_utilization}%, Memory Used: {memory_used}MB, Memory Total: {memory_total}MB, Memory Utilization: {memory_utilization:.2f}%")
|
| 73 |
+
return gpu_status
|
| 74 |
+
|
| 75 |
+
except Exception as e:
|
| 76 |
+
print(f"Error getting GPU stats: {e}")
|
| 77 |
+
|
| 78 |
+
# def update_gpu_status():
|
| 79 |
+
# if torch.cuda.is_available():
|
| 80 |
+
# gpu_info = torch.cuda.get_device_name(0)
|
| 81 |
+
# gpu_memory = torch.cuda.mem_get_info(0)
|
| 82 |
+
# total_memory = gpu_memory[1] / (1024 * 1024)
|
| 83 |
+
# used_memory = (gpu_memory[1] - gpu_memory[0]) / (1024 * 1024)
|
| 84 |
+
|
| 85 |
+
# gpu_status = f"GPU: {gpu_info}\nTotal Memory: {total_memory:.2f} MB\nUsed Memory: {used_memory:.2f} MB"
|
| 86 |
+
# else:
|
| 87 |
+
# gpu_status = "No GPU available"
|
| 88 |
+
# return gpu_status
|
| 89 |
+
|
| 90 |
+
def update_cpu_status():
|
| 91 |
+
import datetime
|
| 92 |
+
# Get the current time
|
| 93 |
+
current_time = datetime.datetime.now().time()
|
| 94 |
+
# Convert the time to a string
|
| 95 |
+
time_str = current_time.strftime("%H:%M:%S")
|
| 96 |
+
|
| 97 |
+
cpu_percent = psutil.cpu_percent()
|
| 98 |
+
cpu_status = f"CPU Usage: {cpu_percent}% {time_str}"
|
| 99 |
+
return cpu_status
|
| 100 |
+
|
| 101 |
+
def update_status():
|
| 102 |
+
gpu_status = update_gpu_status()
|
| 103 |
+
cpu_status = update_cpu_status()
|
| 104 |
+
return gpu_status, cpu_status
|
| 105 |
+
|
| 106 |
+
def refresh_status():
|
| 107 |
+
return update_status()
|
| 108 |
+
|
| 109 |
demo = gr.Blocks()
|
| 110 |
|
| 111 |
mf_transcribe = gr.Interface(
|
|
|
|
| 117 |
label="Model Name",
|
| 118 |
value=DEFAULT_MODEL_NAME,
|
| 119 |
placeholder="Enter the model name",
|
| 120 |
+
info="Some available models: distil-whisper/distil-large-v3 distil-whisper/distil-medium.en Systran/faster-distil-whisper-large-v3 Systran/faster-whisper-large-v3 Systran/faster-whisper-medium openai/whisper-tiny, openai/whisper-base, openai/whisper-medium, openai/whisper-large-v3",
|
| 121 |
),
|
| 122 |
],
|
| 123 |
outputs=[gr.TextArea(label="Transcription"), gr.TextArea(label="Transcription Info")],
|
|
|
|
| 139 |
label="Model Name",
|
| 140 |
value=DEFAULT_MODEL_NAME,
|
| 141 |
placeholder="Enter the model name",
|
| 142 |
+
info="Some available models: openai/whisper-tiny, openai/whisper-base, openai/whisper-medium, openai/whisper-large-v2",
|
| 143 |
),
|
| 144 |
],
|
| 145 |
outputs=[gr.TextArea(label="Transcription"), gr.TextArea(label="Transcription Info")],
|
|
|
|
| 151 |
),
|
| 152 |
allow_flagging="never",
|
| 153 |
)
|
|
|
|
| 154 |
with demo:
|
| 155 |
gr.TabbedInterface([mf_transcribe, file_transcribe], ["Microphone", "Audio file"])
|
| 156 |
+
|
| 157 |
+
with gr.Row():
|
| 158 |
+
refresh_button = gr.Button("Refresh Status") # Create a refresh button
|
| 159 |
+
|
| 160 |
+
gpu_status_output = gr.Textbox(label="GPU Status", interactive=False)
|
| 161 |
+
cpu_status_output = gr.Textbox(label="CPU Status", interactive=False)
|
| 162 |
+
|
| 163 |
+
# Link the refresh button to the refresh_status function
|
| 164 |
+
refresh_button.click(refresh_status, None, [gpu_status_output, cpu_status_output])
|
| 165 |
+
|
| 166 |
+
# Load the initial status using update_status function
|
| 167 |
+
demo.load(update_status, inputs=None, outputs=[gpu_status_output, cpu_status_output], every=2, queue=False)
|
| 168 |
+
|
| 169 |
+
# Launch the Gradio app
|
| 170 |
+
demo.launch(share=True)
|
requirements.txt
CHANGED
|
@@ -1,4 +1,7 @@
|
|
| 1 |
git+https://github.com/huggingface/transformers
|
| 2 |
torch
|
| 3 |
yt-dlp
|
| 4 |
-
gradio==4.8.0
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
git+https://github.com/huggingface/transformers
|
| 2 |
torch
|
| 3 |
yt-dlp
|
| 4 |
+
gradio==4.8.0
|
| 5 |
+
gpustat
|
| 6 |
+
moviepy
|
| 7 |
+
freeze
|