Skip to content

readme_whisper_exec_mac.md

Latest commit

 

History

History
102 lines (75 loc) · 3.53 KB

readme_whisper_exec_mac.md

File metadata and controls

102 lines (75 loc) · 3.53 KB

OpenAi's Whisper as Standalone Mac CLI

Compatibility:

  • Platform: MacOS-14.1-arm64-arm-64bit (M2)
  • Python 3.11.5
  • pip 23.3.1
  • setuptools 68.2.2
  • ffmpeg 6.1.0

Dev Guide | Way to work with it:

  • Clone it. git clone git@github.com:TheLanguageArchive/whisper_standalone_executable.git --depth 1

  • cd whisper_standalone_executable/whisper

  • mkdir -p docs/mac/logs/ ; conda list > docs/mac/logs/transcriber_cli_env_mac_base.txt

  • Create Anaconda (Conda) Environment : transcriber_cli_env_mac_202312211128

  • conda create --name transcriber_cli_env_mac_202312211128 python=3.11.5 -y

  • Select the created Python Environment in the PyCharm / Terminal

    • conda activate transcriber_cli_env_mac_202312211128
    • mkdir -p docs/mac/logs/ ; conda list > docs/mac/logs/transcriber_cli_env_mac_202312211128_init.txt

  • Install Pre-req Dependencies:

    • pip install -r requirements.txt
    • conda install -c conda-forge ffmpeg=6.1.0 -y

  • mkdir -p docs/mac/logs/ ; conda list > docs/mac/logs/transcriber_cli_env_mac_202312211128_final.txt

Clean

rm -r build &&
rm -r dist &&
rm -r whisper_transcriptions &&
rm transcriber_cli_mac_001202312211131.spec

Running as python script

From whisper project root folder.

python cli.py ./../transcriber_jar/src/test/resources/videos/testFile202312061352.mp4 --fp16=False --model=large-v3 --language=en --output_format=json --output_dir=./whisper_transcriptions/ --word_timestamps=True

Packaging for Release(deliverable):

Create the executable(s) file in dist directory that can be run from the command line:

Sample command to run from OSx

Notes:

  • osPathsSeparator :
  • versionWithoutDots
pyinstaller cli.py \
  --noconsole \
  --onefile \
  --clean \
  --log-level=DEBUG \
  --argv-emulation \
  --name transcriber_cli_mac_001202312211131 \
  --add-data "whisper/assets/*":whisper/assets \
  --add-binary ~/anaconda3/envs/transcriber_cli_env_mac_202312211128/bin/ffmpeg:bin \
  --add-binary ~/anaconda3/envs/transcriber_cli_env_mac_202312211128/bin/ffprobe:bin

OR pre-pend following line to the *.spec file import sys ; sys.setrecursionlimit(sys.getrecursionlimit() * 5)

And then

pyinstaller --clean --log-level=DEBUG transcriber_cli_mac_001202312211131.spec

For testing from the project root directory transcriber_cli execute the following command:

"dist/transcriber_cli_mac_001202312211131" ./../transcriber_jar/src/test/resources/videos/testFile202312061352.mp4 --fp16=False --model=large-v3 --language=en --output_format=json --output_dir=./whisper_transcriptions/ --word_timestamps=True

Distribution

mkdir -p ./../transcriber_jar/src/main/resources/releases/ ; \
cp "./dist/transcriber_cli_mac_001202312211131" "./../transcriber_jar/src/main/resources/releases/transcriber_cli_mac_001202312211131" 

conda deactivate

In whisper directory

"./../transcriber_jar/src/main/resources/releases/transcriber_cli_mac_001202312211131" ./../transcriber_jar/src/test/resources/videos/testFile202312061352.mp4 --fp16=False --model=large-v3 --language=en --output_format=json --output_dir=./whisper_transcriptions/ --word_timestamps=True

Transcription files will be created.

NOTE: Ideally, we should be able to assume that after installation if this is the first file copied and ran. It should work. While the executable tries to package everything it needs there are still some libraries which are expected to be present in the host system.

Potentially useful info

Cache location: ~/.cache/whisper/