-
Notifications
You must be signed in to change notification settings - Fork 445
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
1 parent
24ebcea
commit ae0b2e0
Showing
5 changed files
with
201 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,2 +1,4 @@ | ||
zipformer_transducer | ||
paraformer | ||
zipformer_ctc | ||
zipformer_ctc_hlg |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,11 @@ | ||
# Introduction | ||
|
||
This folder contains examples about using sherpa-onnx's object pascal | ||
APIs with streaming models for speech recognition. | ||
|
||
|File|Description| | ||
|----|-----------| | ||
|./run-paraformer.sh|Use a streaming Paraformer model for speech recognition| | ||
|./run-zipformer-ctc-hlg.sh|Use a streaming Zipformer CTC model for speech recognition| | ||
|./run-zipformer-ctc.sh|Use a streaming Zipformer CTC model with HLG for speech recognition| | ||
|./run-zipformer-transducer.sh|Use a Zipformer transducer model for speech recognition| |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,89 @@ | ||
{ Copyright (c) 2024 Xiaomi Corporation } | ||
|
||
{ | ||
This file shows how to use a streaming Zipformer CTC model | ||
to decode files. | ||
You can download the model files from | ||
https://github.com/k2-fsa/sherpa-onnx/releases/tag/asr-models | ||
} | ||
|
||
program zipformer_ctc; | ||
|
||
{$mode delphi} | ||
|
||
uses | ||
sherpa_onnx, | ||
DateUtils, | ||
SysUtils; | ||
|
||
|
||
var | ||
Config: TSherpaOnnxOnlineRecognizerConfig; | ||
Recognizer: TSherpaOnnxOnlineRecognizer; | ||
Stream: TSherpaOnnxOnlineStream; | ||
RecognitionResult: TSherpaOnnxOnlineRecognizerResult; | ||
Wave: TSherpaOnnxWave; | ||
WaveFilename: AnsiString; | ||
TailPaddings: array of Single; | ||
|
||
Start: TDateTime; | ||
Stop: TDateTime; | ||
|
||
Elapsed: Single; | ||
Duration: Single; | ||
RealTimeFactor: Single; | ||
begin | ||
Initialize(Config); | ||
|
||
{Please visit https://github.com/k2-fsa/sherpa-onnx/releases/tag/asr-models | ||
to download model files used in this file.} | ||
Config.ModelConfig.Zipformer2Ctc.Model := './sherpa-onnx-streaming-zipformer-ctc-small-2024-03-18/ctc-epoch-30-avg-3-chunk-16-left-128.int8.onnx'; | ||
Config.ModelConfig.Tokens := './sherpa-onnx-streaming-zipformer-ctc-small-2024-03-18/tokens.txt'; | ||
Config.ModelConfig.Provider := 'cpu'; | ||
Config.ModelConfig.NumThreads := 1; | ||
Config.ModelConfig.Debug := False; | ||
|
||
WaveFilename := './sherpa-onnx-streaming-zipformer-ctc-small-2024-03-18/test_wavs/8k.wav'; | ||
|
||
Wave := SherpaOnnxReadWave(WaveFilename); | ||
|
||
Recognizer := TSherpaOnnxOnlineRecognizer.Create(Config); | ||
|
||
Start := Now; | ||
|
||
Stream := Recognizer.CreateStream(); | ||
|
||
|
||
Stream.AcceptWaveform(Wave.Samples, Wave.SampleRate); | ||
|
||
SetLength(TailPaddings, Round(Wave.SampleRate * 0.5)); {0.5 seconds of padding} | ||
Stream.AcceptWaveform(TailPaddings, Wave.SampleRate); | ||
|
||
Stream.InputFinished(); | ||
|
||
while Recognizer.IsReady(Stream) do | ||
Recognizer.Decode(Stream); | ||
|
||
RecognitionResult := Recognizer.GetResult(Stream); | ||
|
||
Stop := Now; | ||
|
||
Elapsed := MilliSecondsBetween(Stop, Start) / 1000; | ||
Duration := Length(Wave.Samples) / Wave.SampleRate; | ||
RealTimeFactor := Elapsed / Duration; | ||
|
||
WriteLn(RecognitionResult.ToString); | ||
WriteLn(Format('NumThreads %d', [Config.ModelConfig.NumThreads])); | ||
WriteLn(Format('Elapsed %.3f s', [Elapsed])); | ||
WriteLn(Format('Wave duration %.3f s', [Duration])); | ||
WriteLn(Format('RTF = %.3f/%.3f = %.3f', [Elapsed, Duration, RealTimeFactor])); | ||
|
||
{Free resources to avoid memory leak. | ||
Note: You don't need to invoke them for this simple script. | ||
However, you have to invoke them in your own large/complex project. | ||
} | ||
FreeAndNil(Stream); | ||
FreeAndNil(Recognizer); | ||
end. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,90 @@ | ||
{ Copyright (c) 2024 Xiaomi Corporation } | ||
|
||
{ | ||
This file shows how to use a streaming Zipformer CTC model | ||
with HLG to decode files. | ||
You can download the model files from | ||
https://github.com/k2-fsa/sherpa-onnx/releases/tag/asr-models | ||
} | ||
|
||
program zipformer_ctc_hlg; | ||
|
||
{$mode delphi} | ||
|
||
uses | ||
sherpa_onnx, | ||
DateUtils, | ||
SysUtils; | ||
|
||
|
||
var | ||
Config: TSherpaOnnxOnlineRecognizerConfig; | ||
Recognizer: TSherpaOnnxOnlineRecognizer; | ||
Stream: TSherpaOnnxOnlineStream; | ||
RecognitionResult: TSherpaOnnxOnlineRecognizerResult; | ||
Wave: TSherpaOnnxWave; | ||
WaveFilename: AnsiString; | ||
TailPaddings: array of Single; | ||
|
||
Start: TDateTime; | ||
Stop: TDateTime; | ||
|
||
Elapsed: Single; | ||
Duration: Single; | ||
RealTimeFactor: Single; | ||
begin | ||
Initialize(Config); | ||
|
||
{Please visit https://github.com/k2-fsa/sherpa-onnx/releases/tag/asr-models | ||
to download model files used in this file.} | ||
Config.ModelConfig.Zipformer2Ctc.Model := './sherpa-onnx-streaming-zipformer-ctc-small-2024-03-18/ctc-epoch-30-avg-3-chunk-16-left-128.int8.onnx'; | ||
Config.ModelConfig.Tokens := './sherpa-onnx-streaming-zipformer-ctc-small-2024-03-18/tokens.txt'; | ||
Config.ModelConfig.Provider := 'cpu'; | ||
Config.ModelConfig.NumThreads := 1; | ||
Config.ModelConfig.Debug := True; | ||
Config.CtcFstDecoderConfig.Graph := './sherpa-onnx-streaming-zipformer-ctc-small-2024-03-18/HLG.fst'; | ||
|
||
WaveFilename := './sherpa-onnx-streaming-zipformer-ctc-small-2024-03-18/test_wavs/8k.wav'; | ||
|
||
Wave := SherpaOnnxReadWave(WaveFilename); | ||
|
||
Recognizer := TSherpaOnnxOnlineRecognizer.Create(Config); | ||
|
||
Start := Now; | ||
|
||
Stream := Recognizer.CreateStream(); | ||
|
||
|
||
Stream.AcceptWaveform(Wave.Samples, Wave.SampleRate); | ||
|
||
SetLength(TailPaddings, Round(Wave.SampleRate * 0.5)); {0.5 seconds of padding} | ||
Stream.AcceptWaveform(TailPaddings, Wave.SampleRate); | ||
|
||
Stream.InputFinished(); | ||
|
||
while Recognizer.IsReady(Stream) do | ||
Recognizer.Decode(Stream); | ||
|
||
RecognitionResult := Recognizer.GetResult(Stream); | ||
|
||
Stop := Now; | ||
|
||
Elapsed := MilliSecondsBetween(Stop, Start) / 1000; | ||
Duration := Length(Wave.Samples) / Wave.SampleRate; | ||
RealTimeFactor := Elapsed / Duration; | ||
|
||
WriteLn(RecognitionResult.ToString); | ||
WriteLn(Format('NumThreads %d', [Config.ModelConfig.NumThreads])); | ||
WriteLn(Format('Elapsed %.3f s', [Elapsed])); | ||
WriteLn(Format('Wave duration %.3f s', [Duration])); | ||
WriteLn(Format('RTF = %.3f/%.3f = %.3f', [Elapsed, Duration, RealTimeFactor])); | ||
|
||
{Free resources to avoid memory leak. | ||
Note: You don't need to invoke them for this simple script. | ||
However, you have to invoke them in your own large/complex project. | ||
} | ||
FreeAndNil(Stream); | ||
FreeAndNil(Recognizer); | ||
end. |