EngineX-Iluvatar/enginex-mr_series-sherpa-onnx
Archived
16
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity
266 Commits 2 Branches 0 Tags
9efe69720df875fde43df3c6467ab4345e043a20
Commit Graph

122 Commits

Author SHA1 Message Date
Fangjun Kuang
d82251cba6 Support the latest offline zipformer model (#180) 2023-06-19 14:09:24 +08:00
Wei Kang
8562711252 Implement context biasing with a Aho Corasick automata (#145) 
* Implement context graph

* Modify the interface to support context biasing

* Support context biasing in modified beam search; add python wrapper

* Support context biasing in python api example

* Minor fixes

* Fix context graph

* Minor fixes

* Fix tests

* Fix style

* Fix style

* Fix comments

* Minor fixes

* Add missing header

* Replace std::shared_ptr with std::unique_ptr for effciency

* Build graph in constructor

* Fix comments

* Minor fixes

* Fix docs
 2023-06-16 14:26:36 +08:00
keanu
1a1b9fd236 RNNLM model support lm_num_thread and lm_provider setting (#173) 
* rnnlm model inference supports num_threads setting

* rnnlm params decouple num_thread and provider with Transducer.

* fix python csrc bug which offline-lm-config.cc and online-lm-config.cc arguments problem

* lm_num_threads and lm_provider set default values

---------

Co-authored-by: cuidongcai1035 <cuidongcai1035@wezhuiyi.com>
 2023-06-12 15:51:27 +08:00
keanu
9c017c2ccb rnnlm model inference supports num_threads setting (#169) 
Co-authored-by: cuidongcai1035 <cuidongcai1035@wezhuiyi.com>
 2023-06-07 09:32:27 +08:00
Jingzhao Ou
fdd49d0538 add batch processing to sherpa-onnx (#166) 2023-06-04 14:13:55 +08:00
Jingzhao Ou
0ed501b8f1 Added provider option to sherpa-onnx and decode-file-c-api (#162) 2023-06-03 04:57:48 +08:00
Fangjun Kuang
d7114da441 Minor fixes (#161) 2023-05-23 15:57:33 +08:00
Fangjun Kuang
44821ae2fb Use fixed decimal point for offline timestamp (#158) 2023-05-22 16:52:38 +08:00
PF Luo
655c619bf3 Fix lm fusion (#157) 
* share GetHypsRowSplits interface and fix getting Topk not taking logprob

* fix lm score of lm fusion and make padding len same with 'icefall/egs/librispeech/ASR/pruned_transducer_stateless7_streaming/decode.py'
 2023-05-15 10:48:45 +08:00
Yuekai Zhang
b8fbf8e5ce Add onnxruntime gpu for cmake (#153) 
* add onnxruntime gpu for cmake

* fix clang

* fix typo

* cpplint
 2023-05-12 22:30:47 +08:00
Fangjun Kuang
cea718e3d8 Support CoreML for macOS (#151) 2023-05-12 15:57:44 +08:00
Jingzhao Ou
0992063de8 Stack and streaming conformer support (#141) 
* added csrc/stack.cc

* stack: added checks

* added copyright info

* passed cpp style checks

* formatted code

* added some support for streaming conformer model support (not verified)

* code lint

* made more progress with streaming conformer support (not working yet)

* passed style check

* changes as suggested by @csukuangfj

* added some debug info

* fixed style check

* Use Cat to replace Stack

* remove debug statements

---------

Co-authored-by: Jingzhao Ou (jou2019) <jou2019@cisco.com>
Co-authored-by: Fangjun Kuang <csukuangfj@gmail.com>
 2023-05-11 14:30:39 +08:00
PF Luo
824b0809a4 add shallow fusion (#147) 2023-05-10 22:30:57 +08:00
cooldoomsday
0bc571f6ee Return timestamp info and tokens in offline ASR 
Co-authored-by: zhangbaofeng@npnets.com <41259@Zbf>
 2023-05-06 10:20:46 +08:00
PF Luo
8c6a6768d5 Add lm rescore to online-modified-beam-search (#133) 2023-05-05 21:23:54 +08:00
PF Luo
aa7108729b share GetHypsRowSplits interface and fix getting Topk not taking logprob (#131) 2023-04-26 11:41:04 +08:00
Fangjun Kuang
86017f9833 Add RNN LM rescore for offline ASR with modified_beam_search (#125) 2023-04-23 17:15:18 +08:00
Fangjun Kuang
ad05f52666 Add timestamps for streaming ASR. (#123) 2023-04-19 16:02:37 +08:00
彭震东
d781fcdeef Use log probs for paraformer (#120) 
* Use log probs for paraformer

* Fix
 2023-04-09 23:28:10 +08:00
Fangjun Kuang
80060c276d Begin to support CTC models (#119) 
Please see https://k2-fsa.github.io/sherpa/onnx/pretrained_models/offline-ctc/nemo/index.html for a list of pre-trained CTC models from NeMo.
 2023-04-07 23:11:34 +08:00
Fangjun Kuang
283e41c557 Fix displaying English words for paraformer models. (#114) 2023-04-04 16:44:37 +08:00
Fangjun Kuang
726680c5e0 Install binaries via pip install (#112) 
When pepole use pip install sherpa-onnx, they also get the following binaries:

(py38) fangjuns-MacBook-Pro:bin fangjun$ ls -lh  sherpa-onnx*
-rwxr-xr-x  1 fangjun  staff    36K Apr  4 13:48 sherpa-onnx
-rwxr-xr-x  1 fangjun  staff    52K Apr  4 13:48 sherpa-onnx-microphone
-rwxr-xr-x  1 fangjun  staff    54K Apr  4 13:48 sherpa-onnx-microphone-offline
-rwxr-xr-x  1 fangjun  staff    37K Apr  4 13:48 sherpa-onnx-offline
-rwxr-xr-x  1 fangjun  staff   634K Apr  4 13:48 sherpa-onnx-offline-websocket-server
-rwxr-xr-x  1 fangjun  staff   710K Apr  4 13:48 sherpa-onnx-online-websocket-client
-rwxr-xr-x  1 fangjun  staff   651K Apr  4 13:48 sherpa-onnx-online-websocket-server
(py38) fangjuns-MacBook-Pro:bin fangjun$ pwd
/Users/fangjun/py38/bin
 2023-04-04 15:45:59 +08:00
Fangjun Kuang
5d3c8edbc9 add python tests (#111) 2023-04-02 23:05:30 +08:00
manyeyes
3f7e0c23ac adding a python api for offline decode (#110) 2023-04-02 13:17:43 +08:00
Fangjun Kuang
f9f9f20d37 Test int8 models (#107) 
* Test int8 models

* Fix displaying help messages

* small fixes

* Fix jni test
 2023-03-31 22:34:10 +08:00
Fangjun Kuang
b535d6db59 Fix building on Linux (#105) 
* Fix building on Linux

* Fix style issues
 2023-03-30 19:56:59 +08:00
Fangjun Kuang
b911915a32 Add microphone support for offline recognizer (#104) 2023-03-30 19:43:05 +08:00
Fangjun Kuang
6707ec4124 add offline websocket server/client (#98) 2023-03-29 21:48:45 +08:00
Jason's Lab
5e5620ea23 add zipformer model_type (#97) 
Co-authored-by: lichao <www.563042811@qq.com>
 2023-03-29 20:36:39 +08:00
Fangjun Kuang
423d89e9a5 Support paraformer. (#95) 2023-03-28 17:59:54 +08:00
Fangjun Kuang
dffb0fd43c Refactor offline recognizer. (#94) 
* Refactor offline recognizer.

The purpose is to make it easier to support different types of models.
 2023-03-27 14:59:40 +08:00
Fangjun Kuang
5572246253 Add non-streaming ASR (#92) 2023-03-26 08:53:42 +08:00
Fangjun Kuang
6f92bc7362 Fix issues about DLLs when installing pre-compiled wheels on windows (#91) 2023-03-18 13:55:59 +08:00
Fangjun Kuang
8c6a289e3d Add PackPaddedSequence (#85) 2023-03-08 14:12:20 +08:00
frankyoujian
822e05f169 Support windows x86 platform (#86) 
* support windows x86 platform

* improve cmake

* resolve comments

---------

Co-authored-by: Jian You (jianyou) <jianyou@cisco.com>
 2023-03-07 16:47:47 +08:00
Fangjun Kuang
da5c80cc74 add pad_sequence (#84) 2023-03-05 22:02:50 +08:00
Fangjun Kuang
7cae710751 Support slicing a shallow copy of a 3-d tensor (#83) 2023-03-05 19:18:31 +08:00
Fangjun Kuang
422221b3b4 Add transpose (#82) 2023-03-05 11:27:17 +08:00
Fangjun Kuang
9d8fddef01 Support resampling (#77) 2023-03-03 16:42:33 +08:00
Fangjun Kuang
5f31b22c12 Fix modified beam search for iOS and android (#76) 
* Use Int type for sampling rate

* Fix swift

* Fix iOS
 2023-03-03 15:18:31 +08:00
Fangjun Kuang
7f72c13d9a Code refactoring (#74) 
* Don't reset model state and feature extractor on endpointing

* support passing decoding_method from commandline

* Add modified_beam_search to Python API

* fix C API example

* Fix style issues
 2023-03-03 12:10:59 +08:00
you zou
1e45a9503e fix bugs (#73) 2023-03-02 09:51:36 +08:00
PF Luo
5326d0f81f add modified beam search (#69) 2023-03-01 15:32:54 +08:00
Fangjun Kuang
e0b76655c8 Fix batch decoding for greedy search (#71) 2023-03-01 12:18:20 +08:00
Fangjun Kuang
343e732ccb Refactor python examples (#67) 2023-02-26 20:33:16 +08:00
Fangjun Kuang
5a8c3a6d10 Use deep copy in Clone() (#66) 2023-02-26 14:54:01 +08:00
Fangjun Kuang
475caf22f9 Add iOS support (#65) 2023-02-25 21:56:25 +08:00
Fangjun Kuang
fb1e24bebb Fix endpointing with microphone (#64) 2023-02-25 14:30:44 +08:00
Fangjun Kuang
40522f037b add streaming websocket server and client (#62) 2023-02-24 21:39:51 +08:00
Fangjun Kuang
9064b3f016 Support Android (#59) 2023-02-24 13:57:03 +08:00