mirror of
https://github.com/k2-fsa/icefall.git
synced 2025-09-11 01:54:20 +00:00
Merge branch 'dev_multi_zh-hans' of https://github.com/JinZr/icefall into dev_multi_zh-hans
This commit is contained in:
commit
660e431da1
@ -47,7 +47,7 @@ def get_parser():
|
||||
parser.add_argument(
|
||||
"--training-subset",
|
||||
type=str,
|
||||
default="L",
|
||||
default="train_phase1",
|
||||
help="The training subset for computing fbank feature.",
|
||||
)
|
||||
|
||||
@ -93,7 +93,7 @@ def compute_fbank_kespeech_splits(args):
|
||||
subset = args.training_subset
|
||||
subset = str(subset)
|
||||
num_splits = args.num_splits
|
||||
output_dir = f"data/fbank/{subset}_split_{num_splits}"
|
||||
output_dir = f"data/fbank/KeSpeech/{subset}_split_{num_splits}"
|
||||
output_dir = Path(output_dir)
|
||||
assert output_dir.exists(), f"{output_dir} does not exist!"
|
||||
|
||||
|
@ -47,13 +47,13 @@ def has_no_oov(
|
||||
|
||||
def preprocess_kespeech():
|
||||
src_dir = Path("data/manifests/KeSpeech")
|
||||
output_dir = Path("data/fbank")
|
||||
output_dir = Path("data/fbank/KeSpeech")
|
||||
output_dir.mkdir(exist_ok=True)
|
||||
|
||||
# Note: By default, we preprocess all sub-parts.
|
||||
# You can delete those that you don't need.
|
||||
# For instance, if you don't want to use the L subpart, just remove
|
||||
# the line below containing "L"
|
||||
# For instance, if you don't want to use the test subpart, just remove
|
||||
# the line below containing "test"
|
||||
dataset_parts = (
|
||||
"dev_phase1",
|
||||
"dev_phase2",
|
||||
|
Loading…
x
Reference in New Issue
Block a user