Fangjun Kuang
Fangjun Kuang
It wastes me more than 3 hours to figure out the real reason why the upload fails. Hope this pull-request can avoid wasting other's time. The code is from https://github.com/Anaconda-Platform/anaconda-client/issues/501#issuecomment-470742898...
The TCN model and the max-pooling loss are basically the same as the one used inside Mobvoi. Also, one of the contributors did his internship at Mobvoi. I would recommend...
### 先决条件 (Prerequisites) - [X] 我已确认这个镜像源从未在 [其他 issues](https://github.com/tuna/issues) 中讨论过。 I am sure that this repo has NEVER been discussed in [other issues](https://github.com/tuna/issues). - [X] 我已确认这个镜像源没有我选择的镜像站上线。 I am sure that this...
`libtorch` is easy to use. However, the size of its shared libraries is large, see below.  It is nice to support the following frameworks that are more lightweight compared...
Will first make it work for the streaming model [ConvEmformer2](https://github.com/k2-fsa/icefall/tree/master/egs/librispeech/ASR/conv_emformer_transducer_stateless2)
Here are some initial results. For the following test wav (Note: Its name should be `1089-134686-0002.wav`, not `1089-134686-0001.wav`) https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/blob/main/test_wavs/1089-134686-0001.wav The ground truth word timestamp from https://github.com/CorentinJ/librispeech-alignments is ``` ",AFTER,EARLY,NIGHTFALL,THE,YELLOW,LAMPS,WOULD,LIGHT,UP,HERE,AND,THERE,THE,SQUALID,QUARTER,OF,THE,BROTHELS," "0.360,0.730,1.040,1.770,1.900,2.160,2.590,2.760,3.070,3.270,3.520,3.660,4.090,4.210,4.780,5.310,5.420,5.500,6.160,6.625"...
群号码为 744602236 若使用中有什么问题，或者有什么建议，都可以在群里面或者 github 上面提出来. See also https://github.com/k2-fsa/icefall/issues/498
FYI: We have created a huggingface space https://huggingface.co/spaces/k2-fsa/automatic-speech-recognition that uses pre-trained models from [icefall](https://github.com/k2-fsa/icefall) together with sherpa for automatic speech recognition. You can try it using your browser, either uploading...