Commit Graph

  • a39aa8a59d scripts updated jinzr 2024-03-13 10:16:35 +08:00
  • c3f6f28116
    Zipformer recipe for Cantonese dataset MDCC (#1537) zr_jin 2024-03-13 10:01:28 +08:00
  • 17b23ae3bd doc str fixed jinzr 2024-03-13 09:46:00 +08:00
  • 9321f8ab7a add librilight yifanyeung 2024-03-13 00:14:20 +08:00
  • 5dd57a5d35 deploy: 81f518ea7c4dc1e709bb10f21aac55dd33712649 csukuangfj 2024-03-12 15:27:36 +00:00
  • 81f518ea7c
    Support different tts model types. (#1541) Fangjun Kuang 2024-03-12 22:29:21 +08:00
  • 4c41443e13 minor fixes Fangjun Kuang 2024-03-12 22:28:01 +08:00
  • 4abfdc7f57 Update doc Fangjun Kuang 2024-03-12 22:25:23 +08:00
  • a92c6df76a minor fixes Fangjun Kuang 2024-03-12 21:59:47 +08:00
  • 7f9cbf1ce2 typo fixes Fangjun Kuang 2024-03-12 21:53:31 +08:00
  • 71e77e0f7c Upate reamd to add a link to a medium model Fangjun Kuang 2024-03-12 16:45:50 +08:00
  • 9681263c0d Update README Fangjun Kuang 2024-03-12 16:39:14 +08:00
  • 750e2ac035 Update prepare.sh jinzr 2024-03-12 14:35:15 +08:00
  • 204a3b2fb2 arg type fixed jinzr 2024-03-12 12:44:26 +08:00
  • 959906e9dc
    Correct alimeeting download link (#1544) BannerWang 2024-03-12 12:44:09 +08:00
  • d887bf8c63 updated scripts for text jinzr 2024-03-12 12:40:44 +08:00
  • d45e4c61e1 Update prepare.sh jinzr 2024-03-12 12:36:52 +08:00
  • a9df06cef4 Update prepare.sh jinzr 2024-03-12 12:34:27 +08:00
  • 9820bf92f6 updated jinzr 2024-03-12 12:24:24 +08:00
  • 4cae6b6c9a text_norm updated jinzr 2024-03-12 12:19:14 +08:00
  • d35cedcd85 text_norm updated jinzr 2024-03-12 12:18:22 +08:00
  • 4a1d4be94a added scripts for char-based lang prep jinzr 2024-03-12 12:12:35 +08:00
  • 6f56309dce add docs for adapter fine-tuning marcoyang 2024-03-12 11:52:56 +08:00
  • 52da82def2 Correct alimeeting download link BannerWang 2024-03-12 11:10:26 +08:00
  • e69b60e579 enable the grad_scale is too small error jinzr 2024-03-11 23:14:14 +08:00
  • 687b9f1c45 fixed formatting issues jinzr 2024-03-11 23:11:04 +08:00
  • ddefabcb7a added scripts jinzr 2024-03-11 23:09:19 +08:00
  • e472fa6840
    fix CutMix init parameter (#1543) jimmy1984xu 2024-03-11 18:37:26 +08:00
  • f0fef41a20 fix CutMix init parameter jimmyxu 2024-03-11 18:34:16 +08:00
  • 3009ba2d5e initial commit marcoyang 2024-03-11 15:17:29 +08:00
  • b33d3820db Support different tts model types. Fangjun Kuang 2024-03-11 12:42:43 +08:00
  • 6806810666 bug fix marcoyang 2024-03-11 12:18:03 +08:00
  • 3492d9415c add lora version of ActivationDropouAndLinear; currently a simple version marcoyang 2024-03-11 12:02:35 +08:00
  • a421792863 added scripts for streaming related stuff jinzr 2024-03-11 11:08:15 +08:00
  • b2d1975f0e init commit jinzr 2024-03-11 11:04:33 +08:00
  • 60691efddf added RESULTS.md jinzr 2024-03-11 10:59:32 +08:00
  • bb8f6b0ef7 add lora to the in_proj of the feedforward module marcoyang 2024-03-11 10:45:36 +08:00
  • 78b39d9a7d added scripts for exporting model jinzr 2024-03-11 09:38:42 +08:00
  • 660f647886
    Merge branch 'k2-fsa:master' into k2ssl Yifan Yang 2024-03-10 13:10:36 +08:00
  • 8b7ca604c3 minor updates jinzr 2024-03-09 14:09:58 +08:00
  • f62fc7f05f Create encoder_interface.py jinzr 2024-03-09 12:55:42 +08:00
  • 3371c8cff1 init commit for the zipformer recipe jinzr 2024-03-09 12:52:39 +08:00
  • 59837bea09 added manifest statistics jinzr 2024-03-08 20:53:56 +08:00
  • 60986c3ac1
    Fix default value for --context-size in icefall. (#1538) Fangjun Kuang 2024-03-08 20:47:13 +08:00
  • 0e2ee0a9f2 Update requirements.txt jinzr 2024-03-08 20:45:05 +08:00
  • 8ae987225d Fix default value for --context-size in icefall. Fangjun Kuang 2024-03-08 20:40:04 +08:00
  • ced6593179 Create prepare_lang_fst.py jinzr 2024-03-08 19:36:41 +08:00
  • 7f1cb9e991 misc. fixes jinzr 2024-03-08 19:34:06 +08:00
  • 123cd9cb6d Update preprocess_mdcc.py jinzr 2024-03-08 19:32:07 +08:00
  • 27c9a90dc6 misc. fixes jinzr 2024-03-08 19:28:47 +08:00
  • 7d22fef6f2 Update preprocess_mdcc.py jinzr 2024-03-08 19:25:13 +08:00
  • 336944c9b3 Update preprocess_mdcc.py jinzr 2024-03-08 19:22:31 +08:00
  • 8b96efe9e5 misc. fix jinzr 2024-03-08 19:21:17 +08:00
  • c8dbc01308 Update preprocess_mdcc.py jinzr 2024-03-08 19:20:21 +08:00
  • ba01616103 Update preprocess_mdcc.py jinzr 2024-03-08 19:18:19 +08:00
  • 1c509a24f8 handle code switching cases jinzr 2024-03-08 19:14:09 +08:00
  • bca12ed23c Update prepare.sh jinzr 2024-03-08 17:02:57 +08:00
  • 0abb1c346f update mod to executable jinzr 2024-03-08 16:57:34 +08:00
  • 814b59d78d Create README.md jinzr 2024-03-08 16:51:57 +08:00
  • 821ec9db13 init commit jinzr 2024-03-08 16:49:43 +08:00
  • ae61bd4090
    Minor fixes for the commonvoice recipe (#1534) zr_jin 2024-03-08 11:01:11 +08:00
  • 5272a71ec9 also add lora in SelfAttention (for the value proj) marcoyang 2024-03-08 10:40:18 +08:00
  • 89632bf748 Update train.py jinzr 2024-03-08 02:56:53 +08:00
  • ff53cb0458 minor fixes jinzr 2024-03-08 02:10:57 +08:00
  • 4e90233785 fix for issue https://github.com/k2-fsa/icefall/issues/1531 jinzr 2024-03-07 19:33:13 +08:00
  • af62891812 init commit jinzr 2024-03-07 19:27:54 +08:00
  • 5df24c1685
    Whisper large fine-tuning on wenetspeech, mutli-hans-zh (#1483) Yuekai Zhang 2024-03-07 18:04:27 +07:00
  • cdb3fb5675
    add text norm script for pl (#1532) zr_jin 2024-03-07 18:47:29 +08:00
  • 2c54d6b357 remove submodule Yuekai Zhang 2024-03-07 17:46:36 +08:00
  • ab57bb52e0
    fixed a formatting issue zr_jin 2024-03-07 17:25:00 +08:00
  • 21af721ca1
    Update train.py zr_jin 2024-03-07 16:49:33 +08:00
  • e96f533b44
    Update compute_fbank_wenetspeech_splits.py zr_jin 2024-03-07 16:48:29 +08:00
  • 262792f821
    Update prepare.sh zr_jin 2024-03-07 16:46:37 +08:00
  • 211ce4c5fd
    Update train.py zr_jin 2024-03-07 16:45:48 +08:00
  • 4cca65ac98
    Update compute_fbank_wenetspeech_splits.py zr_jin 2024-03-07 16:43:47 +08:00
  • 1c6a6a2011
    Update train.py zr_jin 2024-03-07 16:35:41 +08:00
  • 9bc1ad87b4 add script for decoding marcoyang 2024-03-07 15:42:33 +08:00
  • 19e21ba3ff
    Merge branch 'master' into whisper_zh Yuekai Zhang 2024-03-07 13:53:35 +07:00
  • a00c0c5279 add speechio results Yuekai Zhang 2024-03-07 14:44:38 +08:00
  • a56034799b first version of lora, could just run marcoyang 2024-03-07 12:28:39 +08:00
  • 5ffc976afc Update preprocess_commonvoice.py jinzr 2024-03-07 01:20:35 +08:00
  • b422e7a97f add speechio Yuekai Zhang 2024-03-06 22:09:01 +08:00
  • 87b19c0bd2 add finetune.py marcoyang 2024-03-06 17:06:40 +08:00
  • 3b3256a426 add files again marcoyang 2024-03-06 17:02:21 +08:00
  • 3c92bdb5fa add more files marcoyang 2024-03-06 12:28:20 +08:00
  • 335a9962de
    Fixed formatting issue of PR #1528 (#1530) zr_jin 2024-03-06 08:43:45 +08:00
  • 6953980d4f Update train.py jinzr 2024-03-06 02:24:49 +08:00
  • 86a51e0771 initial commit marcoyang 2024-03-05 18:30:43 +08:00
  • b7f36b40a1
    Update train.py Rezakh20 2024-03-05 12:23:10 +03:30
  • bf5787c438
    Update train.py Rezakh20 2024-03-05 12:21:30 +03:30
  • ff430b465f
    Add num_features to train.py for training WSASR (#1528) Rezakh20 2024-03-05 12:10:30 +03:30
  • 50b575a2f1 load checkpoint from specific path Yuekai Zhang 2024-03-05 16:37:29 +08:00
  • 8836e004c8
    Update egs/librispeech/WSASR/conformer_ctc2/train.py Rezakh20 2024-03-05 11:41:51 +03:30
  • b1be695a89
    Update egs/librispeech/WSASR/conformer_ctc2/train.py Rezakh20 2024-03-05 11:27:08 +03:30
  • 9ee3d0bdf0
    Update egs/librispeech/WSASR/conformer_ctc2/train.py Rezakh20 2024-03-05 11:26:58 +03:30
  • 2dc2112b20
    Update egs/librispeech/WSASR/conformer_ctc2/train.py Rezakh20 2024-03-05 11:26:28 +03:30
  • bc01398902
    Add num_features arg Rezakh20 2024-03-05 11:10:06 +03:30
  • 41a4648eb7
    Rename num_features to num-features in args Rezakh20 2024-03-05 09:42:35 +03:30
  • 67435077c1
    Update train.py Rezakh20 2024-03-05 09:40:41 +03:30
  • 648495d555
    Update train.py Rezakh20 2024-03-05 09:38:19 +03:30