| | |
|---|
| | | |
|---|
| | | |
|---|
| | | def prepare_data(args, distributed_option): |
|---|
| | | if args.dataset_type == "small" and args.train_data_path_and_name_and_type is not None: |
|---|
| | | return |
|---|
| | | if args.dataset_type == "large" and args.train_data_file is not None: |
|---|
| | | return |
|---|
| | | distributed = distributed_option.distributed |
|---|
| | | if not hasattr(args, "train_set"): |
|---|
| | | args.train_set = "train" |
|---|
| | | if not hasattr(args, "dev_set"): |
|---|
| | | args.dev_set = "validation" |
|---|
| | | if not distributed or distributed_option.dist_rank == 0: |
|---|
| | | filter_wav_text(args.data_dir, args.train_set) |
|---|
| | | filter_wav_text(args.data_dir, args.dev_set) |
|---|
