From b15db52e4e67da8a133a67e8ffa415386de48b40 Mon Sep 17 00:00:00 2001
From: zhuyunfeng <10596244@qq.com>
Date: 星期二, 09 五月 2023 23:03:15 +0800
Subject: [PATCH] Add contributor
---
funasr/bin/asr_inference.py | 191 +++++++----------------------------------------
1 files changed, 31 insertions(+), 160 deletions(-)
diff --git a/funasr/bin/asr_inference.py b/funasr/bin/asr_inference.py
index 985ff50..a52e94a 100644
--- a/funasr/bin/asr_inference.py
+++ b/funasr/bin/asr_inference.py
@@ -41,23 +41,19 @@
from funasr.utils.types import str_or_none
from funasr.utils import asr_utils, wav_utils, postprocess_utils
from funasr.models.frontend.wav_frontend import WavFrontend
+from funasr.tasks.asr import frontend_choices
header_colors = '\033[95m'
end_colors = '\033[0m'
-global_asr_language: str = 'zh-cn'
-global_sample_rate: Union[int, Dict[Any, int]] = {
- 'audio_fs': 16000,
- 'model_fs': 16000
-}
class Speech2Text:
"""Speech2Text class
Examples:
>>> import soundfile
- >>> speech2text = Speech2Text("asr_config.yml", "asr.pth")
+ >>> speech2text = Speech2Text("asr_config.yml", "asr.pb")
>>> audio, rate = soundfile.read("speech.wav")
>>> speech2text(audio)
[(text, token, token_int, hypothesis object), ...]
@@ -97,7 +93,11 @@
)
frontend = None
if asr_train_args.frontend is not None and asr_train_args.frontend_conf is not None:
- frontend = WavFrontend(cmvn_file=cmvn_file, **asr_train_args.frontend_conf)
+ if asr_train_args.frontend=='wav_frontend':
+ frontend = WavFrontend(cmvn_file=cmvn_file, **asr_train_args.frontend_conf)
+ else:
+ frontend_class=frontend_choices.get_class(asr_train_args.frontend)
+ frontend = frontend_class(**asr_train_args.frontend_conf).eval()
logging.info("asr_model: {}".format(asr_model))
logging.info("asr_train_args: {}".format(asr_train_args))
@@ -116,7 +116,7 @@
# 2. Build Language model
if lm_train_config is not None:
lm, lm_train_args = LMTask.build_model_from_file(
- lm_train_config, lm_file, device
+ lm_train_config, lm_file, None, device
)
scorers["lm"] = lm.lm
@@ -198,7 +198,7 @@
"""
assert check_argument_types()
-
+
# Input as audio signal
if isinstance(speech, np.ndarray):
speech = torch.tensor(speech)
@@ -256,142 +256,6 @@
assert check_return_type(results)
return results
-
-# def inference(
-# maxlenratio: float,
-# minlenratio: float,
-# batch_size: int,
-# beam_size: int,
-# ngpu: int,
-# ctc_weight: float,
-# lm_weight: float,
-# penalty: float,
-# log_level: Union[int, str],
-# data_path_and_name_and_type,
-# asr_train_config: Optional[str],
-# asr_model_file: Optional[str],
-# cmvn_file: Optional[str] = None,
-# lm_train_config: Optional[str] = None,
-# lm_file: Optional[str] = None,
-# token_type: Optional[str] = None,
-# key_file: Optional[str] = None,
-# word_lm_train_config: Optional[str] = None,
-# bpemodel: Optional[str] = None,
-# allow_variable_data_keys: bool = False,
-# streaming: bool = False,
-# output_dir: Optional[str] = None,
-# dtype: str = "float32",
-# seed: int = 0,
-# ngram_weight: float = 0.9,
-# nbest: int = 1,
-# num_workers: int = 1,
-# **kwargs,
-# ):
-# assert check_argument_types()
-# if batch_size > 1:
-# raise NotImplementedError("batch decoding is not implemented")
-# if word_lm_train_config is not None:
-# raise NotImplementedError("Word LM is not implemented")
-# if ngpu > 1:
-# raise NotImplementedError("only single GPU decoding is supported")
-#
-# logging.basicConfig(
-# level=log_level,
-# format="%(asctime)s (%(module)s:%(lineno)d) %(levelname)s: %(message)s",
-# )
-#
-# if ngpu >= 1 and torch.cuda.is_available():
-# device = "cuda"
-# else:
-# device = "cpu"
-#
-# # 1. Set random-seed
-# set_all_random_seed(seed)
-#
-# # 2. Build speech2text
-# speech2text_kwargs = dict(
-# asr_train_config=asr_train_config,
-# asr_model_file=asr_model_file,
-# cmvn_file=cmvn_file,
-# lm_train_config=lm_train_config,
-# lm_file=lm_file,
-# token_type=token_type,
-# bpemodel=bpemodel,
-# device=device,
-# maxlenratio=maxlenratio,
-# minlenratio=minlenratio,
-# dtype=dtype,
-# beam_size=beam_size,
-# ctc_weight=ctc_weight,
-# lm_weight=lm_weight,
-# ngram_weight=ngram_weight,
-# penalty=penalty,
-# nbest=nbest,
-# streaming=streaming,
-# )
-# logging.info("speech2text_kwargs: {}".format(speech2text_kwargs))
-# speech2text = Speech2Text(**speech2text_kwargs)
-#
-# # 3. Build data-iterator
-# loader = ASRTask.build_streaming_iterator(
-# data_path_and_name_and_type,
-# dtype=dtype,
-# batch_size=batch_size,
-# key_file=key_file,
-# num_workers=num_workers,
-# preprocess_fn=ASRTask.build_preprocess_fn(speech2text.asr_train_args, False),
-# collate_fn=ASRTask.build_collate_fn(speech2text.asr_train_args, False),
-# allow_variable_data_keys=allow_variable_data_keys,
-# inference=True,
-# )
-#
-# finish_count = 0
-# file_count = 1
-# # 7 .Start for-loop
-# # FIXME(kamo): The output format should be discussed about
-# asr_result_list = []
-# if output_dir is not None:
-# writer = DatadirWriter(output_dir)
-# else:
-# writer = None
-#
-# for keys, batch in loader:
-# assert isinstance(batch, dict), type(batch)
-# assert all(isinstance(s, str) for s in keys), keys
-# _bs = len(next(iter(batch.values())))
-# assert len(keys) == _bs, f"{len(keys)} != {_bs}"
-# #batch = {k: v[0] for k, v in batch.items() if not k.endswith("_lengths")}
-#
-# # N-best list of (text, token, token_int, hyp_object)
-# try:
-# results = speech2text(**batch)
-# except TooShortUttError as e:
-# logging.warning(f"Utterance {keys} {e}")
-# hyp = Hypothesis(score=0.0, scores={}, states={}, yseq=[])
-# results = [[" ", ["<space>"], [2], hyp]] * nbest
-#
-# # Only supporting batch_size==1
-# key = keys[0]
-# for n, (text, token, token_int, hyp) in zip(range(1, nbest + 1), results):
-# # Create a directory: outdir/{n}best_recog
-# if writer is not None:
-# ibest_writer = writer[f"{n}best_recog"]
-#
-# # Write the result to each file
-# ibest_writer["token"][key] = " ".join(token)
-# ibest_writer["token_int"][key] = " ".join(map(str, token_int))
-# ibest_writer["score"][key] = str(hyp.score)
-#
-# if text is not None:
-# text_postprocessed = postprocess_utils.sentence_postprocess(token)
-# item = {'key': key, 'value': text_postprocessed}
-# asr_result_list.append(item)
-# finish_count += 1
-# asr_utils.print_progress(finish_count / file_count)
-# if writer is not None:
-# ibest_writer["text"][key] = text
-# return asr_result_list
-
def inference(
maxlenratio: float,
minlenratio: float,
@@ -421,6 +285,7 @@
ngram_weight: float = 0.9,
nbest: int = 1,
num_workers: int = 1,
+ mc: bool = False,
**kwargs,
):
inference_pipeline = inference_modelscope(
@@ -451,6 +316,7 @@
ngram_weight=ngram_weight,
nbest=nbest,
num_workers=num_workers,
+ mc=mc,
**kwargs,
)
return inference_pipeline(data_path_and_name_and_type, raw_inputs)
@@ -483,9 +349,13 @@
ngram_weight: float = 0.9,
nbest: int = 1,
num_workers: int = 1,
+ mc: bool = False,
+ param_dict: dict = None,
**kwargs,
):
assert check_argument_types()
+ ncpu = kwargs.get("ncpu", 1)
+ torch.set_num_threads(ncpu)
if batch_size > 1:
raise NotImplementedError("batch decoding is not implemented")
if word_lm_train_config is not None:
@@ -493,6 +363,9 @@
if ngpu > 1:
raise NotImplementedError("only single GPU decoding is supported")
+ for handler in logging.root.handlers[:]:
+ logging.root.removeHandler(handler)
+
logging.basicConfig(
level=log_level,
format="%(asctime)s (%(module)s:%(lineno)d) %(levelname)s: %(message)s",
@@ -533,6 +406,9 @@
def _forward(data_path_and_name_and_type,
raw_inputs: Union[np.ndarray, torch.Tensor] = None,
output_dir_v2: Optional[str] = None,
+ fs: dict = None,
+ param_dict: dict = None,
+ **kwargs,
):
# 3. Build data-iterator
if data_path_and_name_and_type is None and raw_inputs is not None:
@@ -542,6 +418,8 @@
loader = ASRTask.build_streaming_iterator(
data_path_and_name_and_type,
dtype=dtype,
+ fs=fs,
+ mc=mc,
batch_size=batch_size,
key_file=key_file,
num_workers=num_workers,
@@ -550,7 +428,7 @@
allow_variable_data_keys=allow_variable_data_keys,
inference=True,
)
-
+
finish_count = 0
file_count = 1
# 7 .Start for-loop
@@ -575,7 +453,7 @@
except TooShortUttError as e:
logging.warning(f"Utterance {keys} {e}")
hyp = Hypothesis(score=0.0, scores={}, states={}, yseq=[])
- results = [[" ", ["<space>"], [2], hyp]] * nbest
+ results = [[" ", ["sil"], [2], hyp]] * nbest
# Only supporting batch_size==1
key = keys[0]
@@ -586,30 +464,23 @@
# Write the result to each file
ibest_writer["token"][key] = " ".join(token)
- # ibest_writer["token_int"][key] = " ".join(map(str, token_int))
+ ibest_writer["token_int"][key] = " ".join(map(str, token_int))
ibest_writer["score"][key] = str(hyp.score)
if text is not None:
- text_postprocessed = postprocess_utils.sentence_postprocess(token)
+ text_postprocessed, _ = postprocess_utils.sentence_postprocess(token)
item = {'key': key, 'value': text_postprocessed}
asr_result_list.append(item)
finish_count += 1
asr_utils.print_progress(finish_count / file_count)
if writer is not None:
ibest_writer["text"][key] = text
+
+ logging.info("uttid: {}".format(key))
+ logging.info("text predictions: {}\n".format(text))
return asr_result_list
return _forward
-
-def set_parameters(language: str = None,
- sample_rate: Union[int, Dict[Any, int]] = None):
- if language is not None:
- global global_asr_language
- global_asr_language = language
- if sample_rate is not None:
- global global_sample_rate
- global_sample_rate = sample_rate
-
def get_parser():
parser = config_argparse.ArgumentParser(
@@ -781,4 +652,4 @@
if __name__ == "__main__":
- main()
+ main()
\ No newline at end of file
--
Gitblit v1.9.1