| | |
| | | @contextmanager |
| | | def autocast(enabled=True): |
| | | yield |
| | | from funasr.utils.load_utils import load_audio_and_text_image_video, extract_fbank |
| | | from funasr.utils.load_utils import load_audio_text_image_video, extract_fbank |
| | | from funasr.utils import postprocess_utils |
| | | from funasr.utils.datadir_writer import DatadirWriter |
| | | |
| | |
| | | |
| | | # extract fbank feats |
| | | time1 = time.perf_counter() |
| | | audio_sample_list = load_audio_and_text_image_video(data_in, fs=frontend.fs, audio_fs=kwargs.get("fs", 16000)) |
| | | audio_sample_list = load_audio_text_image_video(data_in, fs=frontend.fs, audio_fs=kwargs.get("fs", 16000)) |
| | | time2 = time.perf_counter() |
| | | meta_data["load_data"] = f"{time2 - time1:0.3f}" |
| | | speech, speech_lengths = extract_fbank(audio_sample_list, data_type=kwargs.get("data_type", "sound"), |
| | |
| | | meta_data[ |
| | | "batch_data_time"] = speech_lengths.sum().item() * frontend.frame_shift * frontend.lfr_n / 1000 |
| | | |
| | | speech.to(device=kwargs["device"]), speech_lengths.to(device=kwargs["device"]) |
| | | speech = speech.to(device=kwargs["device"]) |
| | | speech_lengths = speech_lengths.to(device=kwargs["device"]) |
| | | |
| | | # hotword |
| | | self.hotword_list = self.generate_hotwords_list(kwargs.get("hotword", None), tokenizer=tokenizer, frontend=frontend) |