774 Commits

Author SHA1 Message Date
yaozengwei
20207f0e4e filter inf loss in ctc_loss 2023-02-15 22:22:21 +08:00
yaozengwei
8e893576af fix reduction conformer_ctc3/train.py 2023-02-14 20:38:10 +08:00
yaozengwei
dace53b125 add ctc_decode.py 2023-02-13 21:21:30 +08:00
yaozengwei
251ff62880 Merge remote-tracking branch 'k2-fsa/master' into ctc-latency-rnnt 2023-02-13 20:08:58 +08:00
Zengwei Yao
25ee50e27c
add ctc-greedy-search with timestamps (#905) 2023-02-13 19:45:09 +08:00
yaozengwei
9eaff3364c add ctc-greedy-search with timestamps 2023-02-13 19:10:33 +08:00
Desh Raj
6a8b649e56
Add small streaming Zipformer transducer model (#903) 2023-02-13 15:53:28 +08:00
Yifan Yang
c34ee67691
Update generate_model_from_checkpoint.py (#901) 2023-02-13 14:05:38 +08:00
Fangjun Kuang
c102e7fbf0
more fixes for lstm3 to support exporting to ncnn (#902) 2023-02-13 12:16:43 +08:00
yaozengwei
3d4b8eb58b add ctc branch, with delay penalty 2023-02-13 12:11:35 +08:00
Fangjun Kuang
48c2c22dbe
Fix export to ncnn for lstm3 (#900) 2023-02-13 11:44:25 +08:00
yaozengwei
6ce36934cb init files 2023-02-13 11:17:30 +08:00
KajiMaCN
57604aac34
fix tal_csasr data pre-processing (#898) 2023-02-10 21:28:19 +08:00
xiabingquan
cba6ecc1d1
Update README.md (#894) 2023-02-09 23:54:45 +08:00
emilyluj
59ac8bfc70
fix mmi graph compiler bug. (#895) 2023-02-09 18:32:03 +08:00
Yifan Yang
5cd1636cb3
Fix a bug in decode.py (#893)
Co-authored-by: yifanyang <yifanyeung@yifanyangs-MacBook-Pro.local>
2023-02-09 12:12:23 +08:00
Fangjun Kuang
e916027bfe
Fix doc typos for onnx export (#891) 2023-02-09 10:33:40 +08:00
Karel Vesely
35e5a2475c
Librispeech, validate_manifest.py (#890) 2023-02-09 07:57:02 +08:00
Fangjun Kuang
2b995639b7
Add ONNX support for Zipformer and ConvEmformer (#884) 2023-02-09 00:02:38 +08:00
Zengwei Yao
af735eb75b
Get alignments using lhotse workflows align-with-torchaudio (#888)
* add lhotse workflow align-with-torchaudio

* modify related decode.py files
2023-02-08 21:54:35 +08:00
Zengwei Yao
d12e6f098c
Get (start, end) timestamps for CTC models (#876)
* parse timestamps and texts for BPE-based models

* parse timestamps (frame indexes) and texts for other cases

* add test functions

* add parse_fsa_timestamps_and_texts function, test in conformer_ctc3/decode.py

* calculate symbol delay for (start, end) timestamps
2023-02-07 21:43:16 +08:00
yaozengwei
fa740be117 minor change 2023-02-07 21:42:02 +08:00
yaozengwei
cca3113817 calculate symbol delay for (start, end) timestamps 2023-02-07 21:16:56 +08:00
Fangjun Kuang
7ae03f6c88
Add onnx export support for pruned_transducer_stateless5 (#883) 2023-02-07 17:47:08 +08:00
Yifan Yang
ffbf6d9199
Add generate_averaged_model.py (#882) 2023-02-07 16:19:08 +08:00
Fangjun Kuang
8d3810e289
Simplify ONNX export (#881)
* Simplify ONNX export

* Fix ONNX CI tests
2023-02-07 15:01:59 +08:00
yaozengwei
3e1d14b9f8 add parse_fsa_timestamps_and_texts function, test in conformer_ctc3/decode.py 2023-02-07 14:59:29 +08:00
Fangjun Kuang
52f3a747be
Refactor onnx export for streaming zipformer (#879) 2023-02-07 12:12:26 +08:00
yaozengwei
0e4f7c59c2 Merge remote-tracking branch 'k2-fsa/master' into ctc-timestamp 2023-02-07 10:30:24 +08:00
Zengwei Yao
5a05b95730
add params.hlg_scale (#880) 2023-02-06 23:21:46 +08:00
yaozengwei
c0804166ea add more test cases 2023-02-06 20:32:57 +08:00
yaozengwei
0526e8fcd8 add test functions 2023-02-06 19:37:53 +08:00
yaozengwei
00e6a8dd78 parse timestamps (frame indexes) and texts for other cases 2023-02-06 17:09:52 +08:00
Yifan Yang
caf23546ed
No more T < S after frame_reducer (#875)
* No more T < S after frame_reducer

* Fix for style check

* Adjust the permissions

* Add support for inference to frame_reducer

* Fix for flake8 check

---------

Co-authored-by: yifanyang <yifanyeung@yifanyangs-MacBook-Pro.local>
2023-02-06 12:17:45 +08:00
Yuekai Zhang
bf5f0342a2
Add streaming onnx export for zipformer (#831)
* add streaming onnx export for zipformer

* update triton support

* add comments

* add ci test

* add onnxmltools for fp16 onnx export
2023-02-06 10:37:07 +08:00
yaozengwei
1a6a035d87 parse timestamps and texts for BPE-based models 2023-02-05 18:20:49 +08:00
Yifan Yang
029c8566e4
Small fix for frame_reducer.py (#871) 2023-02-03 17:49:54 +08:00
yaozengwei
dd0047e605 Merge remote-tracking branch 'k2-fsa/master' 2023-02-03 15:06:54 +08:00
Yifan Yang
bffce413f0
Fix filename ctc_guild_decode_bs.py -> ctc_guide_decode_bs.py (#870)
* fix filename ctc_guild_decode_bs.py -> ctc_guide_decode_bs.py

---------

Co-authored-by: yifanyang <yifanyeung@yifanyangs-MacBook-Pro.local>
2023-02-03 12:32:06 +08:00
Zengwei Yao
1e6d6f8160
shuffle full Librispeech for zipformer recipes (#869)
* shuffle libri
2023-02-03 11:54:57 +08:00
Yifan Yang
e36ea89112
update result.md for pruned_transducer_stateless7_ctc_bs (#865) 2023-02-01 21:04:56 +08:00
Yifan Yang
d8234e199c
Add export to ONNX for Zipformer+CTC using blank skip (#861)
* Add export to ONNX for Zipformer+CTC using blank skip

---------

Co-authored-by: yifanyang <yifanyeung@yifanyangs-MacBook-Pro.local>
2023-01-31 15:57:03 +08:00
BuaaAlban
e9019511eb
Fix bug in streaming_conformer_ctc egs (#862)
* Update train.py

Fix transducer lstm egs bug as mentioned in issue 579

* Update train.py

fix dataloader bug
2023-01-31 15:19:50 +08:00
Yifan Yang
e277e31e37
update huggingface link of zipformer_ctc_blankskip.rst (#858)
* update huggingface link

* update link

---------

Co-authored-by: yifanyang <yifanyeung@yifanyangs-MacBook-Pro.local>
2023-01-29 15:35:36 +08:00
Meng Wei
74a2069f94
fix expired links (#856) 2023-01-28 14:43:47 +08:00
Teo Wen Shen
1ce2bc1ee0
edit comments (#852) 2023-01-28 13:47:21 +08:00
Zengwei Yao
6b1ab71dc9
hardcode --filter-uneven-sized-batch (#854) 2023-01-27 21:24:12 +08:00
Wei Kang
f5ff7a18eb
Fix the unclear description for streaming model (#849) 2023-01-17 11:28:59 +08:00
Fangjun Kuang
0af3e7beda
fix export for stateless4 (#844) 2023-01-16 20:26:36 +08:00
Zengwei Yao
2a463a420d
Filter uneven-sized batch (#843)
* add filter_uneven_sized_batch fucntion

* set --filter-uneven-sized-batch=True as default
2023-01-16 20:15:35 +08:00