Support non-streaming zipformer CTC ASR models (#2340)

This PR adds support for non-streaming Zipformer CTC ASR models across 
multiple language bindings, WebAssembly, examples, and CI workflows.

- Introduces a new OfflineZipformerCtcModelConfig in C/C++, Python, Swift, Java, Kotlin, Go, Dart, Pascal, and C# APIs
- Updates initialization, freeing, and recognition logic to include Zipformer CTC in WASM and Node.js
- Adds example scripts and CI steps for downloading, building, and running Zipformer CTC models

Model doc is available at
https://k2-fsa.github.io/sherpa/onnx/pretrained_models/offline-ctc/icefall/zipformer.html
This commit is contained in:
Fangjun Kuang
2025-07-04 15:57:07 +08:00
committed by GitHub
parent ef16455cb5
commit 3bf986d08d
71 changed files with 2121 additions and 68 deletions

View File

@@ -154,6 +154,23 @@ rm sherpa-onnx-dolphin-base-ctc-multi-lang-int8-2025-04-02.tar.bz2
node ./test-offline-dolphin-ctc.js
```
## ./test-offline-zipformer-ctc.js
[./test-offline-zipformer-ctc.js](./test-offline-zipformer-ctc.js) demonstrates
how to decode a file with a Zipformer CTC model. In the code we use
[sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03](https://k2-fsa.github.io/sherpa/onnx/pretrained_models/offline-ctc/icefall/zipformer.html#sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03-chinese).
You can use the following command to run it:
```bash
wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03.tar.bz2
tar xvf sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03.tar.bz2
rm sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03.tar.bz2
node ./test-offline-zipformer-ctc.js
```
## ./test-offline-nemo-ctc.js
[./test-offline-nemo-ctc.js](./test-offline-nemo-ctc.js) demonstrates

View File

@@ -0,0 +1,35 @@
// Copyright (c) 2025 Xiaomi Corporation (authors: Fangjun Kuang)
//
const fs = require('fs');
const {Readable} = require('stream');
const wav = require('wav');
const sherpa_onnx = require('sherpa-onnx');
function createOfflineRecognizer() {
let config = {
modelConfig: {
zipformerCtc: {
model: './sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03/model.int8.onnx',
},
tokens: './sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03/tokens.txt',
}
};
return sherpa_onnx.createOfflineRecognizer(config);
}
const recognizer = createOfflineRecognizer();
const stream = recognizer.createStream();
const waveFilename =
'./sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03/test_wavs/0.wav';
const wave = sherpa_onnx.readWave(waveFilename);
stream.acceptWaveform(wave.sampleRate, wave.samples);
recognizer.decode(stream);
const text = recognizer.getResult(stream).text;
console.log(text);
stream.free();
recognizer.free();