speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

add paraformer online opt infer code

Browse Source
This commit is contained in:
haoneng.lhn 2023-09-14 12:25:44 +08:00
parent 14d0ba3a10
commit f41a1276ff
1 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
funasr/bin/asr_inference_launch.py
View File

@ -953,14 +953,15 @@ def inference_paraformer_online(
# FIXME(kamo): The output format should be discussed about
raw_inputs = torch.unsqueeze(raw_inputs, axis=0)
asr_result_list = []
cache = _prepare_cache(cache, chunk_size=chunk_size, batch_size=1)
cache = _prepare_cache(cache, chunk_size=chunk_size, encoder_chunk_look_back=encoder_chunk_look_back,
decoder_chunk_look_back=decoder_chunk_look_back, batch_size=1)
item = {}
if data_path_and_name_and_type is not None and data_path_and_name_and_type[2] == "sound":
sample_offset = 0
speech_length = raw_inputs.shape[1]
stride_size = chunk_size[1] * 960
cache = _prepare_cache(cache, chunk_size=chunk_size, batch_size=1,
encoder_chunk_look_back=encoder_chunk_look_back, decoder_chunk_look_back=decoder_chunk_look_back)
cache = _prepare_cache(cache, chunk_size=chunk_size, encoder_chunk_look_back=encoder_chunk_look_back,
decoder_chunk_look_back=decoder_chunk_look_back, batch_size=1)
final_result = ""
for sample_offset in range(0, speech_length, min(stride_size, speech_length - sample_offset)):
if sample_offset + stride_size >= speech_length - 1: