Support streaming zipformer CTC (#496)

* Support streaming zipformer CTC

* test online zipformer2 CTC

* Update doc of sherpa-onnx.cc

* Add Python APIs for streaming zipformer2 ctc

* Add Python API examples for streaming zipformer2 ctc

* Swift API for streaming zipformer2 CTC

* NodeJS API for streaming zipformer2 CTC

* Kotlin API for streaming zipformer2 CTC

* Golang API for streaming zipformer2 CTC

* C# API for streaming zipformer2 CTC

* Release v1.9.6
This commit is contained in:
Fangjun Kuang
2023-12-22 13:46:33 +08:00
committed by GitHub
parent 7634f5f034
commit e475e750ac
70 changed files with 1517 additions and 211 deletions

View File

@@ -5,3 +5,4 @@ tts
vits-vctk
sherpa-onnx-paraformer-zh-2023-09-14
!*.sh
*.bak

View File

@@ -60,6 +60,14 @@ func sherpaOnnxOnlineParaformerModelConfig(
)
}
func sherpaOnnxOnlineZipformer2CtcModelConfig(
model: String = ""
) -> SherpaOnnxOnlineZipformer2CtcModelConfig {
return SherpaOnnxOnlineZipformer2CtcModelConfig(
model: toCPointer(model)
)
}
/// Return an instance of SherpaOnnxOnlineModelConfig.
///
/// Please refer to
@@ -75,6 +83,8 @@ func sherpaOnnxOnlineModelConfig(
tokens: String,
transducer: SherpaOnnxOnlineTransducerModelConfig = sherpaOnnxOnlineTransducerModelConfig(),
paraformer: SherpaOnnxOnlineParaformerModelConfig = sherpaOnnxOnlineParaformerModelConfig(),
zipformer2Ctc: SherpaOnnxOnlineZipformer2CtcModelConfig =
sherpaOnnxOnlineZipformer2CtcModelConfig(),
numThreads: Int = 1,
provider: String = "cpu",
debug: Int = 0,
@@ -83,6 +93,7 @@ func sherpaOnnxOnlineModelConfig(
return SherpaOnnxOnlineModelConfig(
transducer: transducer,
paraformer: paraformer,
zipformer2_ctc: zipformer2Ctc,
tokens: toCPointer(tokens),
num_threads: Int32(numThreads),
provider: toCPointer(provider),

View File

@@ -13,24 +13,47 @@ extension AVAudioPCMBuffer {
}
func run() {
let encoder =
"./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/encoder-epoch-99-avg-1.onnx"
let decoder =
"./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/decoder-epoch-99-avg-1.onnx"
let joiner =
"./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/joiner-epoch-99-avg-1.onnx"
let tokens = "./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/tokens.txt"
var modelConfig: SherpaOnnxOnlineModelConfig
var modelType = "zipformer2-ctc"
var filePath: String
let transducerConfig = sherpaOnnxOnlineTransducerModelConfig(
encoder: encoder,
decoder: decoder,
joiner: joiner
)
modelType = "transducer"
let modelConfig = sherpaOnnxOnlineModelConfig(
tokens: tokens,
transducer: transducerConfig
)
if modelType == "transducer" {
filePath = "./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/test_wavs/1.wav"
let encoder =
"./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/encoder-epoch-99-avg-1.onnx"
let decoder =
"./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/decoder-epoch-99-avg-1.onnx"
let joiner =
"./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/joiner-epoch-99-avg-1.onnx"
let tokens = "./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/tokens.txt"
let transducerConfig = sherpaOnnxOnlineTransducerModelConfig(
encoder: encoder,
decoder: decoder,
joiner: joiner
)
modelConfig = sherpaOnnxOnlineModelConfig(
tokens: tokens,
transducer: transducerConfig
)
} else {
filePath =
"./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13/test_wavs/DEV_T0000000000.wav"
let model =
"./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13/ctc-epoch-20-avg-1-chunk-16-left-128.onnx"
let tokens = "./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13/tokens.txt"
let zipfomer2CtcModelConfig = sherpaOnnxOnlineZipformer2CtcModelConfig(
model: model
)
modelConfig = sherpaOnnxOnlineModelConfig(
tokens: tokens,
zipformer2Ctc: zipfomer2CtcModelConfig
)
}
let featConfig = sherpaOnnxFeatureConfig(
sampleRate: 16000,
@@ -43,7 +66,6 @@ func run() {
let recognizer = SherpaOnnxRecognizer(config: &config)
let filePath = "./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20/test_wavs/1.wav"
let fileURL: NSURL = NSURL(fileURLWithPath: filePath)
let audioFile = try! AVAudioFile(forReading: fileURL as URL)

View File

@@ -20,6 +20,12 @@ if [ ! -d ./sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20 ]; then
rm sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20.tar.bz2
fi
if [ ! -d ./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13 ]; then
wget -q https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13.tar.bz2
tar xvf sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13.tar.bz2
rm sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13.tar.bz2
fi
if [ ! -e ./decode-file ]; then
# Note: We use -lc++ to link against libc++ instead of libstdc++
swiftc \

View File

@@ -22,7 +22,7 @@ if [ ! -d ./sherpa-onnx-whisper-tiny.en ]; then
fi
if [ ! -f ./silero_vad.onnx ]; then
echo "downloading silero_vad"
wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/silero_vad.onnx
wget -q https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/silero_vad.onnx
fi
if [ ! -e ./generate-subtitles ]; then