Fangjun Kuang
Committed by GitHub

Add Chinese models (#16)

@@ -44,7 +44,13 @@ jobs: @@ -44,7 +44,13 @@ jobs:
44 git lfs install 44 git lfs install
45 git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13 45 git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13
46 46
47 - - name: Configure Cmake 47 + - name: Download pretrained model and test-data (Chinese)
  48 + shell: bash
  49 + run: |
  50 + git lfs install
  51 + git clone https://huggingface.co/luomingshuang/icefall_asr_wenetspeech_pruned_transducer_stateless2
  52 +
  53 + - name: Configure CMake
48 shell: bash 54 shell: bash
49 run: | 55 run: |
50 mkdir build 56 mkdir build
@@ -84,3 +90,32 @@ jobs: @@ -84,3 +90,32 @@ jobs:
84 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/onnx/joiner_encoder_proj.onnx \ 90 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/onnx/joiner_encoder_proj.onnx \
85 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/onnx/joiner_decoder_proj.onnx \ 91 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/onnx/joiner_decoder_proj.onnx \
86 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1221-135766-0002.wav 92 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1221-135766-0002.wav
  93 +
  94 + - name: Run tests for ubuntu/macos (Chinese)
  95 + run: |
  96 + time ./build/bin/sherpa-onnx \
  97 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/data/lang_char/tokens.txt \
  98 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/encoder-epoch-10-avg-2.onnx \
  99 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/decoder-epoch-10-avg-2.onnx \
  100 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner-epoch-10-avg-2.onnx \
  101 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_encoder_proj-epoch-10-avg-2.onnx \
  102 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_decoder_proj-epoch-10-avg-2.onnx \
  103 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/test_wavs/DEV_T0000000000.wav
  104 +
  105 + time ./build/bin/sherpa-onnx \
  106 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/data/lang_char/tokens.txt \
  107 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/encoder-epoch-10-avg-2.onnx \
  108 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/decoder-epoch-10-avg-2.onnx \
  109 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner-epoch-10-avg-2.onnx \
  110 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_encoder_proj-epoch-10-avg-2.onnx \
  111 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_decoder_proj-epoch-10-avg-2.onnx \
  112 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/test_wavs/DEV_T0000000001.wav
  113 +
  114 + time ./build/bin/sherpa-onnx \
  115 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/data/lang_char/tokens.txt \
  116 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/encoder-epoch-10-avg-2.onnx \
  117 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/decoder-epoch-10-avg-2.onnx \
  118 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner-epoch-10-avg-2.onnx \
  119 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_encoder_proj-epoch-10-avg-2.onnx \
  120 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_decoder_proj-epoch-10-avg-2.onnx \
  121 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/test_wavs/DEV_T0000000002.wav
@@ -9,7 +9,7 @@ We provide exported models in onnx format and they can be downloaded using @@ -9,7 +9,7 @@ We provide exported models in onnx format and they can be downloaded using
9 the following links: 9 the following links:
10 10
11 - English: <https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13> 11 - English: <https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13>
12 -- Chinese: `TODO` 12 +- Chinese: <https://huggingface.co/luomingshuang/icefall_asr_wenetspeech_pruned_transducer_stateless2>
13 13
14 **NOTE**: We provide only non-streaming models at present. 14 **NOTE**: We provide only non-streaming models at present.
15 15
@@ -17,6 +17,9 @@ the following links: @@ -17,6 +17,9 @@ the following links:
17 **HINT**: The script for exporting the English model can be found at 17 **HINT**: The script for exporting the English model can be found at
18 <https://github.com/k2-fsa/icefall/blob/master/egs/librispeech/ASR/pruned_transducer_stateless3/export.py> 18 <https://github.com/k2-fsa/icefall/blob/master/egs/librispeech/ASR/pruned_transducer_stateless3/export.py>
19 19
  20 +**HINT**: The script for exporting the Chinese model can be found at
  21 +<https://github.com/k2-fsa/icefall/blob/master/egs/wenetspeech/ASR/pruned_transducer_stateless2/export.py>
  22 +
20 # Usage 23 # Usage
21 24
22 ```bash 25 ```bash
@@ -48,3 +51,23 @@ git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-trans @@ -48,3 +51,23 @@ git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-trans
48 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/onnx/joiner_decoder_proj.onnx \ 51 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/onnx/joiner_decoder_proj.onnx \
49 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1089-134686-0001.wav 52 ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1089-134686-0001.wav
50 ``` 53 ```
  54 +
  55 +## Download the pretrained model (Chinese)
  56 +
  57 +**Caution**: You have to run `git lfs install`. Otherwise, you will be **SAD** later.
  58 +
  59 +```bash
  60 +git lfs install
  61 +git clone https://huggingface.co/luomingshuang/icefall_asr_wenetspeech_pruned_transducer_stateless2
  62 +
  63 +./build/bin/sherpa-onnx --help
  64 +
  65 +./build/bin/sherpa-onnx \
  66 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/data/lang_char/tokens.txt \
  67 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/encoder-epoch-10-avg-2.onnx \
  68 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/decoder-epoch-10-avg-2.onnx \
  69 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner-epoch-10-avg-2.onnx \
  70 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_encoder_proj-epoch-10-avg-2.onnx \
  71 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/exp/joiner_decoder_proj-epoch-10-avg-2.onnx \
  72 + ./icefall_asr_wenetspeech_pruned_transducer_stateless2/test_wavs/DEV_T0000000000.wav
  73 +```