init repo

嘉渊

2023-08-05 47343b5c2f4e1256f60f46d8da0aa2e5de39b6c7

 egs/callhome/eend_ola/local/infer.py

@@ -54,7 +54,7 @@
    parser.add_argument(
        "--sampling_rate",
        type=int,
        default=10,
        default=8000,
        help="sampling rate",
    )
    parser.add_argument(
@@ -62,6 +62,12 @@
        type=int,
        default=10,
        help="setting subsampling",
    )
    parser.add_argument(
        "--shuffle",
        type=bool,
        default=True,
        help="shuffle speech in time",
    )
    parser.add_argument(
        "--attractor_threshold",
@@ -98,7 +104,7 @@
    print("Start inference")
    with open(args.output_rttm_file, "w") as wf:
        for wav_id in wav_items.keys():
            print("Process wav: {}\n".format(wav_id))
            print("Process wav: {}".format(wav_id))
            data, rate = sf.read(wav_items[wav_id])
            speech = eend_ola_feature.stft(data, args.frame_size, args.frame_shift)
            speech = eend_ola_feature.transform(speech)