Quentin Fuxa
8c799fa4d1
fix simulstreaming vram leak: cap cross-attn accumulation + token budget
...
fixes #283 , fixes #275
- accumulated_cross_attns was growing unboundedly during decoding loop,
using up to ~5GB for repetition loops. now capped to rolling window of 16
- max_tokens_per_chunk was using TOKENS_PER_SECOND (mel frame rate = 50)
instead of actual text token rate (~15/s), allowing 10-40x too many
decoding steps
- removed unused torch.cat on early return path
- removed dead self.committed/last_result_tokens lists (never read)
- same fixes applied to mlx variant
2026-02-11 22:10:00 +01:00
Quentin Fuxa
8923337380
fix --direct-english-translation not setting task=translate for localagreement backends
...
the flag was only used for tokenizer language selection but never
actually passed to whisper/faster-whisper transcribe calls. also init
OpenaiApiASR.task and read from transcribe_kargs.
fixes #306
2026-02-11 22:10:00 +01:00
Quentin Fuxa
aded1649ae
fix model_cache_dir + direct_english_translation task in simulstreaming
...
pass actual cache dir instead of None, and use proper task string
instead of boolean for AlignAttConfig
fixes #310
2026-02-11 22:10:00 +01:00
Quentin Fuxa
3b535e857a
fix NoneType concatenation in add_translation
...
fixes #296
2026-02-11 22:10:00 +01:00
Quentin Fuxa
d649250b9a
fix Segment classmethod call + isinstance type narrowing
...
fixes #331 , fixes #329
2026-02-11 22:10:00 +01:00
Quentin Fuxa
7735478286
add insert_audio_chunk to DiartDiarization
...
fixes #332
2026-02-11 22:10:00 +01:00
Quentin Fuxa
b9e72d2b9a
add probability field to ASRToken
...
fixes #330 , fixes #313
2026-02-11 22:10:00 +01:00
Emmanuel Schmidbauer
d59ddbaeae
Fix critical thread safety issues
2026-01-09 11:23:19 -05:00
Anton Jacobson
3db5d81a20
update diarization lag after stream analysed
2025-12-18 14:13:28 +01:00
Quentin Fuxa
ec09898a9f
fixes #301
2025-12-06 10:19:50 +01:00
Quentin Fuxa
719e8b1a20
adapt online for mlx detection
2024-11-25 23:52:00 +01:00
Quentin Fuxa
f1b47178d8
adapt online for mlx detection
2024-11-25 23:52:00 +01:00
Quentin Fuxa
59db08e961
loader for full mlx
2024-11-25 23:52:00 +01:00
Quentin Fuxa
6fc20b9562
new dec class
2024-11-21 23:52:00 +01:00
Quentin Fuxa
fac8659161
uses native mlx function for attention
2024-11-21 23:52:00 +01:00
Quentin Fuxa
4d9332ce7d
fixes #299
2025-12-05 17:54:14 +01:00
Quentin Fuxa
62444ce746
session parameter required in OnnxWrapper
2025-12-05 15:37:18 +01:00
Quentin Fuxa
2431a6bf91
isolated VAD states per user: .onnx: share a stateless model. .jit: require duplicating the model.
...
Co-authored-by: eschmidbauer <eschmidbauer@gmail.com >
2025-12-05 15:27:14 +01:00
Zizheng Guo
30ddd522a4
Fix local agreement backend, removing excess parameter, fixes https://github.com/QuentinFuxa/WhisperLiveKit/issues/295
2025-12-04 16:45:23 +08:00
Quentin Fuxa
82cd24bb75
LoRa path v0 - functional
2025-11-29 17:21:10 +01:00
Quentin Fuxa
d45c397c6a
simulstreaming: limit n tokens to prevent hallucinations
2025-11-28 21:41:19 +01:00
Quentin Fuxa
1d88ba9d69
Fixes #294 . improve model path backend detection and file extraction
2025-11-27 23:14:00 +01:00
Quentin Fuxa
c0965c6c31
Lines to Segments. Merging dataclasses
2025-11-27 21:54:58 +01:00
Quentin Fuxa
7faa21f95f
alignatt: enable model sharing by removing hooks and centralizing session state. Solves #282
...
Co-authored-by: Emmanuel Schmidbauer <eschmidbauer@gmail.com >
2025-11-25 23:07:42 +01:00
Quentin Fuxa
4e9f951551
correct silences handling when language not auto
2025-11-20 11:20:00 +01:00
Quentin Fuxa
870141298c
isort
2025-11-23 11:20:00 +01:00
Quentin Fuxa
a175d1a327
fixes silence detected but never reported by silero
2025-11-23 11:20:00 +01:00
Quentin Fuxa
6206fff118
0.2.15
2025-11-21 23:52:00 +01:00
Quentin Fuxa
b5067249c0
stt/diar/nllw alignment: internal rework 5
2025-11-20 23:52:00 +01:00
Quentin Fuxa
f4f9831d39
stt/diar/nllw alignment: internal rework 5
2025-11-20 23:52:00 +01:00
Quentin Fuxa
254faaf64c
stt/diar/nllw alignment: internal rework 5
2025-11-20 23:52:00 +01:00
Quentin Fuxa
8e7aea4fcf
internal rework 4
2025-11-20 23:45:20 +01:00
Quentin Fuxa
270faf2069
internal rework 3
2025-11-20 22:28:30 +01:00
Quentin Fuxa
b7c1cc77cc
internal rework 2
2025-11-20 22:06:38 +01:00
Quentin Fuxa
9a45ec221c
internal rework 1
2025-11-20 12:58:38 +01:00
Quentin Fuxa
b7d20a0ff0
segment attribution in result formatter
2025-11-19 21:10:28 +01:00
Quentin Fuxa
c1bb9c2bde
reduce flickering remaining_time_transcription
2025-11-19 19:09:37 +01:00
Quentin Fuxa
11e9def0b2
diarization corrections
2025-11-19 19:06:03 +01:00
Quentin Fuxa
3104f40f6e
fixes #279 #278
2025-11-19 18:17:50 +01:00
Quentin Fuxa
e9b4ceeee5
Add audio partial silence in chunks handling. bump to 0.2.14.post3
2025-11-17 22:52:00 +01:00
Quentin Fuxa
437641fb43
reduce min-chunk-size to 0.1, set default model to base
2027-04-25 23:52:00 +02:00
Quentin Fuxa
bfd60b3921
Add audio partial silence in chunks handling. bump to 0.2.14.post2
2025-11-17 22:52:00 +01:00
Quentin Fuxa
1e67bf97f0
improve buffering when use of heavy models
2027-04-25 23:52:00 +02:00
Quentin Fuxa
bbd4fd6cff
Merge branch 'improve_EOS_handling'
2025-11-16 22:30:31 +01:00
Quentin Fuxa
28985962a0
Silence handling: finish transcription even if not validated at the BEGINNING of the silence
2025-11-16 22:29:08 +01:00
Quentin Fuxa
a38c103fcd
simulstreaming coreml encoder compatibility
2025-11-16 21:24:14 +01:00
Quentin Fuxa
4d2ffb24f8
coreml conversion
2025-11-16 19:11:43 +01:00
Quentin Fuxa
1bbbb7903c
lora loader in shared whisper core
2025-11-16 18:44:35 +01:00
Quentin Fuxa
80b77998f9
Refactor backend handling
2025-11-15 19:51:41 +01:00
Quentin Fuxa
d310f7e25f
hf compatibility
2025-11-15 18:34:19 +01:00
