update docs

speech_asr

2023-02-14 cb82e9fdef0f2cb5b80fda4eaf9c2ef202934191

 funasr/bin/asr_inference_paraformer_vad_punc.py

@@ -40,18 +40,10 @@
from funasr.tasks.vad import VADTask
from funasr.utils.timestamp_tools import time_stamp_lfr6
from funasr.bin.punctuation_infer import Text2Punc
from funasr.torch_utils.forward_adaptor import ForwardAdaptor
from funasr.datasets.preprocessor import CommonPreprocessor
from funasr.punctuation.text_preprocessor import split_to_mini_sentence

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
@@ -548,6 +540,7 @@
    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,
                 ):
        # 3. Build data-iterator
@@ -558,6 +551,7 @@
        loader = ASRTask.build_streaming_iterator(
            data_path_and_name_and_type,
            dtype=dtype,
            fs=fs,
            batch_size=1,
            key_file=key_file,
            num_workers=num_workers,
@@ -659,12 +653,8 @@
                        punc_id_list = ""
                        text_postprocessed_punc = ""

                    item = {'key': key, 'value': text_postprocessed_punc_time_stamp, 'text': text_postprocessed,
                            'time_stamp': time_stamp_postprocessed, 'punc': punc_id_list, 'token': token}
                    if outputs_dict:
                        item = {'text_punc': text_postprocessed_punc, 'text': text_postprocessed,
                                'punc_id': punc_id_list, 'token': token, 'time_stamp': time_stamp_postprocessed}
                        item = {'key': key, 'value': item}
                    item = {'key': key, 'value': text_postprocessed_punc, 'text_postprocessed': text_postprocessed,
                            'time_stamp': time_stamp_postprocessed, 'token': token}
                    asr_result_list.append(item)
                    finish_count += 1
                    # asr_utils.print_progress(finish_count / file_count)