Skip to content

Releases: Purfview/whisper-standalone-win

Faster-Whisper-XXL r192.3.4

27 Mar 01:04
3b611dc
Compare
Choose a tag to compare

Standalone Faster-Whisper implementation using optimized CTranslate2 models.

Includes all Standalone Faster-Whisper features +some additional ones.
Includes all needed libs.
Read there ->#231

Some new stuff in r192.3.4:

New feature: Write intermediate files to temp folder. [ignored on dump]
New feature: Can take JSON files as input to generate subtitles from it according to the settings.
New alternative VAD method : pyannote_v3 [previous same named option renamed to "pyannote_onnx_v3"]
New arg: --nullify_non_speech [for WIP experiments]

Link to the changelog.

Faster-Whisper r192.3

15 Aug 16:04
a07e9de
Compare
Choose a tag to compare

Standalone Faster-Whisper implementation using optimized CTranslate2 models.

GPU execution requires cuBLAS and cuDNN 8.x libs for CUDA v11.x .

Last included commit: #192

Some new stuff in r192.3:

Bugfix: 'one_word' was broken in r192.2.

Link to the changelog.

Whisper-OpenAI r136

30 May 22:49
a07e9de
Compare
Choose a tag to compare

Last included commit: #136.

Notes:

Needs FFmpeg in PATH or place it in the same folder with Whisper executable.
"CPU" build runs only on CPU.

cuBLAS and cuDNN

17 May 21:30
898f61a
Compare
Choose a tag to compare

Place libs in the same folder where Faster-Whisper executable is. Or to:
Windows: To System32 dir.
Linux: To dir in LD_LIBRARY_PATH env.

.7z vs .zip - archives contain same files.

v2 is the last with support for GPUs with Kepler chip.