EngineX-Iluvatar/enginex-mr_series-sherpa-onnx
Archived
16
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Support streaming zipformer CTC (#496)

Browse Source 
* 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
Download Patch File Download Diff File Expand all files Collapse all files

40
python-api-examples/online-decode-files.py
View File

@@ -37,7 +37,20 @@ git lfs pull --include "*.onnx"
./sherpa-onnx-streaming-paraformer-bilingual-zh-en/test_wavs/3.wav \
./sherpa-onnx-streaming-paraformer-bilingual-zh-en/test_wavs/8k.wav
(3) Streaming Conformer CTC from WeNet
(3) Streaming Zipformer2 CTC
wget 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
ls -lh sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13
./python-api-examples/online-decode-files.py \
--zipformer2-ctc=./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13/ctc-epoch-20-avg-1-chunk-16-left-128.onnx \
--tokens=./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13/tokens.txt \
./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13/test_wavs/DEV_T0000000000.wav \
./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13/test_wavs/DEV_T0000000001.wav
(4) Streaming Conformer CTC from WeNet
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/csukuangfj/sherpa-onnx-zh-wenet-wenetspeech
cd sherpa-onnx-zh-wenet-wenetspeech
@@ -51,12 +64,9 @@ git lfs pull --include "*.onnx"
./sherpa-onnx-zh-wenet-wenetspeech/test_wavs/8k.wav
Please refer to
https://k2-fsa.github.io/sherpa/onnx/index.html
and
https://k2-fsa.github.io/sherpa/onnx/pretrained_models/wenet/index.html
to install sherpa-onnx and to download streaming pre-trained models.
https://k2-fsa.github.io/sherpa/onnx/pretrained_models/index.html
to download streaming pre-trained models.
"""
import argparse
import time
@@ -97,6 +107,12 @@ def get_args():
help="Path to the transducer joiner model",
)
parser.add_argument(
"--zipformer2-ctc",
type=str,
help="Path to the zipformer2 ctc model",
)
parser.add_argument(
"--paraformer-encoder",
type=str,
@@ -112,7 +128,7 @@ def get_args():
parser.add_argument(
"--wenet-ctc",
type=str,
help="Path to the wenet ctc model model",
help="Path to the wenet ctc model",
)
parser.add_argument(
@@ -275,6 +291,16 @@ def main():
hotwords_file=args.hotwords_file,
hotwords_score=args.hotwords_score,
)
elif args.zipformer2_ctc:
recognizer = sherpa_onnx.OnlineRecognizer.from_zipformer2_ctc(
tokens=args.tokens,
model=args.zipformer2_ctc,
num_threads=args.num_threads,
provider=args.provider,
sample_rate=16000,
feature_dim=80,
decoding_method="greedy_search",
)
elif args.paraformer_encoder:
recognizer = sherpa_onnx.OnlineRecognizer.from_paraformer(
tokens=args.tokens,

5
python-api-examples/online-websocket-client-decode-file.py
View File

@@ -25,6 +25,7 @@ https://github.com/k2-fsa/sherpa-onnx/blob/master/sherpa-onnx/csrc/online-websoc
import argparse
import asyncio
import json
import logging
import wave
@@ -112,7 +113,7 @@ async def receive_results(socket: websockets.WebSocketServerProtocol):
async for message in socket:
if message != "Done!":
last_message = message
logging.info(message)
logging.info(json.loads(message))
else:
break
return last_message
@@ -151,7 +152,7 @@ async def run(
await websocket.send("Done")
decoding_results = await receive_task
logging.info(f"\nFinal result is:\n{decoding_results}")
logging.info(f"\nFinal result is:\n{json.loads(decoding_results)}")
async def main():

26
python-api-examples/streaming_server.py
View File

@@ -137,6 +137,12 @@ def add_model_args(parser: argparse.ArgumentParser):
help="Path to the transducer joiner model.",
)
parser.add_argument(
"--zipformer2-ctc",
type=str,
help="Path to the model file from zipformer2 ctc",
)
parser.add_argument(
"--wenet-ctc",
type=str,
@@ -405,6 +411,20 @@ def create_recognizer(args) -> sherpa_onnx.OnlineRecognizer:
rule3_min_utterance_length=args.rule3_min_utterance_length,
provider=args.provider,
)
elif args.zipformer2_ctc:
recognizer = sherpa_onnx.OnlineRecognizer.from_zipformer2_ctc(
tokens=args.tokens,
model=args.zipformer2_ctc,
num_threads=args.num_threads,
sample_rate=args.sample_rate,
feature_dim=args.feat_dim,
decoding_method=args.decoding_method,
enable_endpoint_detection=args.use_endpoint != 0,
rule1_min_trailing_silence=args.rule1_min_trailing_silence,
rule2_min_trailing_silence=args.rule2_min_trailing_silence,
rule3_min_utterance_length=args.rule3_min_utterance_length,
provider=args.provider,
)
elif args.wenet_ctc:
recognizer = sherpa_onnx.OnlineRecognizer.from_wenet_ctc(
tokens=args.tokens,
@@ -748,6 +768,8 @@ def check_args(args):
assert args.paraformer_encoder is None, args.paraformer_encoder
assert args.paraformer_decoder is None, args.paraformer_decoder
assert args.zipformer2_ctc is None, args.zipformer2_ctc
assert args.wenet_ctc is None, args.wenet_ctc
elif args.paraformer_encoder:
assert Path(
args.paraformer_encoder
@@ -756,6 +778,10 @@ def check_args(args):
assert Path(
args.paraformer_decoder
).is_file(), f"{args.paraformer_decoder} does not exist"
elif args.zipformer2_ctc:
assert Path(
args.zipformer2_ctc
).is_file(), f"{args.zipformer2_ctc} does not exist"
elif args.wenet_ctc:
assert Path(args.wenet_ctc).is_file(), f"{args.wenet_ctc} does not exist"
else: