Spaces:
Build error
Build error
Commit History
-
Only serve the OpenAI-compatible endpoint with an API key
A server whose UI is protected some other way, such as a proxy that guards only Gradio's own endpoints, would have gained an open transcription endpoint just by updating. openai_api now defaults to null, which serves the endpoint only when an API key is set. true still serves it without a key, and false turns it off even with one.1929c74 -
Add an OpenAI-compatible transcription endpoint
app.py now also serves POST /v1/audio/transcriptions, so clients written for OpenAI or other Whisper servers can transcribe with this app. Gradio 3.38 is kept: launch() returns its FastAPI app, and the route is added to it before blocking. Every option of the Full tab can be sent as an extra form field named as in config.json5, which is how the OpenAI SDKs send extra_body. Options default to the config, except that audio is transcribed as sent (no VAD) and diarization is off. Requests run one at a time outside the UI's queue, delete their upload afterwards, and use their own diarization process, so a UI job can't stop it. The endpoint can require "Authorization: Bearer <key>", set with WHISPER_WEBUI_API_KEY, --api_key or api_key in config.json5, and can be disabled with --openai_api False.06a9401
-
Adding support for setting server root
95b5e37
-
Handle attribute error
dc8bae2
-
Better error handling
6324aa9
-
Adding support for large-v3 in faster-whisper
4749691
-
Make a separate process timeout for diarization
3cd7d59 -
Ensure GPU memory in diarization can be cleaned up
18bb72f -
Adding support for diarization
a1da02d
-
Adding a dummy implementation of Whisper for testing
6250a98
-
Fix unit test
052fe7e
-
Hande NULLs in from_string
ce65f49 -
Use type hints instead
f641aea -
Adding JSON initial prompt
By selecting "json_prompt_mode", you can customize the prompt to each segment. For instance: [ {"segment_index": 0, "prompt": "Hello, how are you?"}, {"segment_index": 1, "prompt": "I'm doing well, how are you?"}, {"segment_index": 2, "prompt": "{0} Fine, thank you.", "format_prompt": true} ]74b7d77
-
[Faster whisper] Add support for large-v1
67b8308
-
Load config file using UTF-8
4a46425
-
Fix preprocessor when words are disabled
ab69310 -
Adding support for word timestamps
f55c594
-
Add inital prompt mode. GITLAB #7
Add a new configuration option `vad_initial_prompt_mode` in config.json5 and the application arguments that have the following two modes: * prepend_all_segments: Preprend the initial prompt to each VAD segment * prepend_first_segment: Only preprend the initial prompt the first VAD segment. This is useful if you're using the prompt to improve the accuracy of the transcription of unusual technical terms consistently throughout a lecture. You can add these terms to the prompt, and then set `vad_initial_prompt_mode` to `prepend_all_segments` to include the prompt in every VAD segment. Note that this will have no effect if you're not using a VAD.8031785
-
Display time taken to run whisperCallable.invoke() (#20)
- Display time taken to run whisperCallable.invoke() (1db1652ddd938383ff52d5b077306586ffab0e4b) Co-authored-by: FlippFuzz <FlippFuzz@users.noreply.huggingface.co>168184d -
Add space to timestamp in fasterWhisperContainer
55b2bd6 -
Add timestamp to verbose output (#21)
- Add timestamp to generated output (5b766a28786398f438626f3d7bfc8738266a038c) Co-authored-by: FlippFuzz <FlippFuzz@users.noreply.huggingface.co>5dcaadd
-
Refactor language list
adca588 -
Fix chinese
e77e696 -
Only print fp16 warning once
a79dd83 -
Support CLI into faster-whisper
33ee1bb
-
Fix WHISPER_IMPLEMENTATION argument
2698c96 -
Fix convert_hf_whisper
9939d13 -
Adding support for faster_whisper
This is a re-implementation of Whisper in CTranslate2 that can be 4x faster and use much less memory than OpenAI's Whisper.295de00
-
Ensure progress bar works for multiple files
c90f138 -
Adding example useage in whisperProgressHook
c963436 -
Support progress for multiple devices
4a9d465 -
Add progress listener to none/VAD
Note that we don't handle progress of parallel transcription yet.33a2c1e
-
Add concurrency count to app.py. Set this to 1 by default.
This is to ensure that the WebUI is accessible from within proxies.d20404a
-
Add more configuration options to config.json5
1acaa19
-
Add configuration file and support for custom models
Custom models can be added to the configuration file, under the "models" section. See the comments for more details.44d964a
-
Do not parallelize period vad
530547e
-
Use correct root dir
7fd072f -
Ensure model is downloaded before spawning sub-processes
7c5d37e -
Add support for multiple input files and output files
cb9ee50 -
Always create N processes for N devices even with small files
That way, we avoid accidentally allocating the same model twice on the same device.479b187
-
Fix 1 CPU in ParallelTranscription
596fd47
-
Add an option for automatically configuring parallel execution
20f75ae
-
Display duration of parallel GPU transcribing
4b698fb -
Support parallel execution of Silero VAD
c0e541b
-
Fix CLI for parallel devices
01fddc0 -
Loop using merged_split instead of devices
9934006 -
Fix chunk splitting
0ec48c0 -
Cache parallel processes
31f7bdb -
Spawn a new process for each GPU
4aa12d0