mirror of
https://github.com/modelscope/FunASR
synced 2025-09-15 14:48:36 +08:00
b454a1054f
* init * update * add LoadConfigFromYaml * update * update * update * del time stat * update * update * update * update * update * update * update * add cpp websocket online 2pass srv * [feature] multithread grpc server * update * update * update * [feature] support 2pass grpc cpp server and python client, can change mode to use offline, online or 2pass decoding * update * update * update * update * add paraformer online onnx model export * add paraformer online onnx model export * add paraformer online onnx model export * add paraformer online onnxruntime * add paraformer online onnxruntime * add paraformer online onnxruntime * fix export paraformer online onnx model bug * for client closed earlier and core dump * support GRPC two pass decoding (#813) * [refator] optimize grpc server pipeline and instruction * [refator] rm useless file * [refator] optimize grpc client pipeline and instruction * [debug] hanlde coredump when client ternimated * [refator] rm useless log * [refator] modify grpc cmake * Create run_server_2pass.sh * Update SDK_tutorial_online_zh.md * Update SDK_tutorial_online.md * Update SDK_advanced_guide_online.md * Update SDK_advanced_guide_online_zh.md * Update SDK_tutorial_online_zh.md * Update SDK_tutorial_online.md * update --------- Co-authored-by: zhaoming <zhaomingwork@qq.com> Co-authored-by: boji123 <boji123@aliyun.com> Co-authored-by: haoneng.lhn <haoneng.lhn@alibaba-inc.com>
31 lines
1.2 KiB
Python
31 lines
1.2 KiB
Python
import soundfile
|
|
from funasr_onnx.paraformer_online_bin import Paraformer
|
|
from pathlib import Path
|
|
|
|
model_dir = "damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online"
|
|
wav_path = '{}/.cache/modelscope/hub/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/example/asr_example.wav'.format(Path.home())
|
|
|
|
chunk_size = [5, 10, 5]
|
|
model = Paraformer(model_dir, batch_size=1, quantize=True, chunk_size=chunk_size, intra_op_num_threads=4) # only support batch_size = 1
|
|
|
|
##online asr
|
|
speech, sample_rate = soundfile.read(wav_path)
|
|
speech_length = speech.shape[0]
|
|
sample_offset = 0
|
|
step = chunk_size[1] * 960
|
|
param_dict = {'cache': dict()}
|
|
final_result = ""
|
|
for sample_offset in range(0, speech_length, min(step, speech_length - sample_offset)):
|
|
if sample_offset + step >= speech_length - 1:
|
|
step = speech_length - sample_offset
|
|
is_final = True
|
|
else:
|
|
is_final = False
|
|
param_dict['is_final'] = is_final
|
|
rec_result = model(audio_in=speech[sample_offset: sample_offset + step],
|
|
param_dict=param_dict)
|
|
if len(rec_result) > 0:
|
|
final_result += rec_result[0]["preds"][0]
|
|
print(rec_result)
|
|
print(final_result)
|