| | |
|---|
| | | import numpy as np |
|---|
| | | import torch |
|---|
| | | import torchaudio |
|---|
| | | import soundfile |
|---|
| | | import yaml |
|---|
| | | from typeguard import check_argument_types |
|---|
| | | |
|---|
| | |
|---|
| | | data_with_index = [(vadsegments[i], i) for i in range(n)] |
|---|
| | | sorted_data = sorted(data_with_index, key=lambda x: x[0][1] - x[0][0]) |
|---|
| | | results_sorted = [] |
|---|
| | | batch_size_token_ms = batch_size_token * 60 |
|---|
| | | |
|---|
| | | batch_size_token_ms = batch_size_token*60 |
|---|
| | | if speech2text.device == "cpu": |
|---|
| | | batch_size_token_ms = 0 |
|---|
| | | batch_size_token_ms = max(batch_size_token_ms, sorted_data[0][0][1] - sorted_data[0][0][0]) |
|---|
| | | |
|---|
| | | batch_size_token_ms_cum = 0 |
|---|
| | | beg_idx = 0 |
|---|
| | | for j, _ in enumerate(range(0, n)): |
|---|
| | |
|---|
| | | raw_inputs = _load_bytes(data_path_and_name_and_type[0]) |
|---|
| | | raw_inputs = torch.tensor(raw_inputs) |
|---|
| | | if data_path_and_name_and_type is not None and data_path_and_name_and_type[2] == "sound": |
|---|
| | | raw_inputs = torchaudio.load(data_path_and_name_and_type[0])[0][0] |
|---|
| | | try: |
|---|
| | | raw_inputs = torchaudio.load(data_path_and_name_and_type[0])[0][0] |
|---|
| | | except: |
|---|
| | | raw_inputs = soundfile.read(data_path_and_name_and_type[0], dtype='float32')[0] |
|---|
| | | if raw_inputs.ndim == 2: |
|---|
| | | raw_inputs = raw_inputs[:, 0] |
|---|
| | | raw_inputs = torch.tensor(raw_inputs) |
|---|
| | | if data_path_and_name_and_type is None and raw_inputs is not None: |
|---|
| | | if isinstance(raw_inputs, np.ndarray): |
|---|
| | | raw_inputs = torch.tensor(raw_inputs) |
|---|
