funasr/models/campplus/model.py
@@ -110,6 +110,7 @@ time2 = time.perf_counter() meta_data["load_data"] = f"{time2 - time1:0.3f}" speech, speech_lengths, speech_times = extract_feature(audio_sample_list) speech = speech.to(device=kwargs["device"]) time3 = time.perf_counter() meta_data["extract_feat"] = f"{time3 - time2:0.3f}" meta_data["batch_data_time"] = np.array(speech_times).sum().item() / 16000.0
