issues
search
huggingface
/
distil-whisper
Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.
MIT License
3.32k
stars
238
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Short form evaluation WER % for Librispeech clean test
#88
guynich
closed
4 months ago
3
Errors running `run_long_form_eval.py`
#87
guynich
closed
3 months ago
1
Training README triggers BuildConfig ValueError
#86
guynich
closed
4 months ago
3
Builderconfig
#85
guynich
closed
4 months ago
0
Whisper-tiny support?
#84
alicewith
opened
4 months ago
2
Cannot do pseudo-labelling due to some error
#83
alicewith
closed
4 months ago
2
Why do we need to tokenized file_id?
#82
macabdul9
closed
3 months ago
1
[training] configure bs in preprocessing
#81
sanchit-gandhi
closed
4 months ago
0
WER Filtering takes too long?
#80
macabdul9
opened
4 months ago
3
correctly mask bos in prompted ids
#79
sanchit-gandhi
closed
5 months ago
0
Avoid preprocessing dataset again?
#78
ghost
closed
4 months ago
1
mask BOS in prompted ids
#77
sanchit-gandhi
closed
5 months ago
1
Bettertransformer is default after transformers 4.36
#76
kamalojasv181
closed
3 months ago
1
Loss Weight ablation experiment
#75
JinYu1998
opened
5 months ago
3
Training / Evaluation fixes
#74
sanchit-gandhi
closed
5 months ago
0
I want to confirm how the knowledge organization is implemented?
#73
hxypqr
opened
5 months ago
1
Is there the "distil-whisper/distil-medium" model?
#72
bk111
opened
5 months ago
1
ddp ERROR
#71
liyingjie1991
opened
5 months ago
1
支持中文吗?
#70
xxm1668
opened
6 months ago
4
The evaluation stuck after unwrap_model during distillation training
#69
HackGiter
closed
5 months ago
1
Very high WER % in extensive benchmark on Fleurs
#68
asusdisciple
closed
5 months ago
3
I can only use its encoder to extract audio features, right? How should I use it? Could you provide an example
#67
wvinzh
opened
6 months ago
1
Update README with speculative decoding resources
#66
sanchit-gandhi
closed
6 months ago
0
What are the settings used for WER calculation in the paper?
#65
hidoba
opened
6 months ago
1
RuntimeError: Error(s) in loading state_dict for WhisperForConditionalGeneration.
#64
Eskaggg
opened
6 months ago
4
High inference time when using chunk size 15
#63
shashikg
opened
6 months ago
2
How can I assign parameters for another language transcribe task?
#62
liuamo
closed
5 months ago
2
Is there a parameter that can keep the numbers in their original language form?
#61
GrahLnn
opened
6 months ago
5
NotImplementedError
#60
wntg
opened
6 months ago
1
Accelerate
#59
wntg
opened
6 months ago
1
Training without real labels
#58
rolisz
opened
6 months ago
1
Can fine tuning for phoneme task?
#57
jackNhat
closed
6 months ago
1
[Training]: Configure pseudo-labels or text labels as targets in KD
#56
sanchit-gandhi
closed
6 months ago
0
ggml-distil-small.en.bin slower than ggml-small.en.bin under whisper.cpp
#55
regularfry
opened
6 months ago
0
[Question] Mixed Speech transcription
#54
Tejaswgupta
opened
6 months ago
5
small uses more memory (but is faster) than medium (ONNX quantized)
#53
kmn1024
closed
6 months ago
5
Fix regex in `get_last_checkpoint`
#52
bofenghuang
closed
6 months ago
1
Fix total_train_steps issue related to gradient_accumulation_steps
#51
bofenghuang
closed
6 months ago
2
Could distil-whisper load local speech dataset?
#50
shuaijiang
opened
6 months ago
7
How to make training data?
#49
satani99
opened
6 months ago
1
.
#48
Tejaswgupta
closed
1 month ago
1
Can we use distil whisper for 50+ concurrent requests on one T4 machine without compromising on latency for each request?
#47
moksh-samespace
opened
7 months ago
1
How can I generate SRT file ?
#46
bmox
closed
6 months ago
4
Why we need KL-divergence
#45
penolove
opened
7 months ago
3
DDP updates
#44
sanchit-gandhi
closed
7 months ago
0
small readme updates
#43
sanchit-gandhi
closed
7 months ago
0
Improve readme training
#42
patrickvonplaten
closed
7 months ago
0
Streaming Implementation of Distil Whisper
#41
souvikqb
closed
7 months ago
2
Does this use FSDP or Deepspeed? Req. for accelerate config
#40
geekyGoku
opened
7 months ago
4
quick readme updates
#39
sanchit-gandhi
closed
7 months ago
0
Previous
Next