From 5b6bd201412636b6d7fa85afaf24e42cd54e52cd Mon Sep 17 00:00:00 2001
From: jmwang66 <wangjiaming.wjm@alibaba-inc.com>
Date: 星期五, 30 十二月 2022 18:50:07 +0800
Subject: [PATCH] Merge branch 'dev' of https://github.com/alibaba-damo-academy/FunASR into dev
---
funasr/bin/asr_inference_paraformer.py | 18 +++++++++---------
1 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/funasr/bin/asr_inference_paraformer.py b/funasr/bin/asr_inference_paraformer.py
index 9dcd0b8..09c61bc 100755
--- a/funasr/bin/asr_inference_paraformer.py
+++ b/funasr/bin/asr_inference_paraformer.py
@@ -92,8 +92,8 @@
if asr_model.frontend is None and frontend_conf is not None:
frontend = WavFrontend(**frontend_conf)
asr_model.frontend = frontend
- logging.info("asr_model: {}".format(asr_model))
- logging.info("asr_train_args: {}".format(asr_train_args))
+ # logging.info("asr_model: {}".format(asr_model))
+ # logging.info("asr_train_args: {}".format(asr_train_args))
asr_model.to(dtype=getattr(torch, dtype)).eval()
ctc = CTCPrefixScorer(ctc=asr_model.ctc, eos=asr_model.eos)
@@ -141,8 +141,8 @@
for scorer in scorers.values():
if isinstance(scorer, torch.nn.Module):
scorer.to(device=device, dtype=getattr(torch, dtype)).eval()
- logging.info(f"Beam_search: {beam_search}")
- logging.info(f"Decoding device={device}, dtype={dtype}")
+ # logging.info(f"Beam_search: {beam_search}")
+ # logging.info(f"Decoding device={device}, dtype={dtype}")
# 5. [Optional] Build Text converter: e.g. bpe-sym -> Text
if token_type is None:
@@ -160,7 +160,7 @@
else:
tokenizer = build_tokenizer(token_type=token_type)
converter = TokenIDConverter(token_list=token_list)
- logging.info(f"Text tokenizer: {tokenizer}")
+ # logging.info(f"Text tokenizer: {tokenizer}")
self.asr_model = asr_model
self.asr_train_args = asr_train_args
@@ -197,9 +197,9 @@
# data: (Nsamples,) -> (1, Nsamples)
# lengths: (1,)
- if len(speech.size()) < 3:
- speech = speech.unsqueeze(0).to(getattr(torch, self.dtype))
- speech_lengths = speech.new_full([1], dtype=torch.long, fill_value=speech.size(1))
+ # if len(speech.size()) < 3:
+ # speech = speech.unsqueeze(0).to(getattr(torch, self.dtype))
+ # speech_lengths = speech.new_full([1], dtype=torch.long, fill_value=speech.size(1))
lfr_factor = max(1, (speech.size()[-1]//80)-1)
batch = {"speech": speech, "speech_lengths": speech_lengths}
@@ -426,7 +426,7 @@
assert len(keys) == _bs, f"{len(keys)} != {_bs}"
# batch = {k: v for k, v in batch.items() if not k.endswith("_lengths")}
- logging.info("decoding, utt_id: {}".format(keys))
+ # logging.info("decoding, utt_id: {}".format(keys))
# N-best list of (text, token, token_int, hyp_object)
time_beg = time.time()
--
Gitblit v1.9.1