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

Merge branch 'dev' of https://github.com/alibaba/FunASR into dev

Browse Source
This commit is contained in:
lzr265946 2022-12-29 17:56:33 +08:00
commit bc3ad9c0f6
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
funasr/bin/asr_inference_uniasr.py
View File

@ -215,14 +215,14 @@ class Speech2Text:
lfr_factor = max(1, (speech.size()[-1] // 80) - 1) lfr_factor = max(1, (speech.size()[-1] // 80) - 1)
# lengths: (1,) # lengths: (1,)
lengths = speech.new_full([1], dtype=torch.long, fill_value=speech.size(1)) lengths = speech.new_full([1], dtype=torch.long, fill_value=speech.size(1))
speech_raw = speech.clone().to(self.device)
if self.frontend is not None: if self.frontend is not None:
feats, feats_len = self.frontend.forward(speech, lengths) feats, feats_len = self.frontend.forward(speech, lengths)
feats = to_device(feats, device=self.device) feats = to_device(feats, device=self.device)
feats_len = feats_len.int() feats_len = feats_len.int()
else: else:
feats = speech_raw feats = speech
feats_len = lengths feats_len = lengths
feats_raw = feats.clone().to(self.device)
batch = {"speech": feats, "speech_lengths": feats_len} batch = {"speech": feats, "speech_lengths": feats_len}
# a. To device # a. To device
@ -235,7 +235,7 @@ class Speech2Text:
if self.decoding_mode == "model1": if self.decoding_mode == "model1":
predictor_outs = self.asr_model.calc_predictor_mask(enc, enc_len) predictor_outs = self.asr_model.calc_predictor_mask(enc, enc_len)
else: else:
enc, enc_len = self.asr_model.encode2(enc, enc_len, feats, feats_len, ind=self.decoding_ind) enc, enc_len = self.asr_model.encode2(enc, enc_len, feats_raw, feats_len, ind=self.decoding_ind)
predictor_outs = self.asr_model.calc_predictor_mask2(enc, enc_len) predictor_outs = self.asr_model.calc_predictor_mask2(enc, enc_len)
scama_mask = predictor_outs[4] scama_mask = predictor_outs[4]