experiments/STT
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Include missing changes in MLS English importer

Browse Source
This commit is contained in:
Reuben Morais 2021-05-04 19:06:18 +02:00
parent ad4025af7d
commit ce71ec0c89
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
bin/import_mls_english.py
View File

@ -29,7 +29,11 @@ def read_ogg_opus_duration(ogg_file_path):
def main(root_dir): def main(root_dir):
for subset in ("test",): for subset in (
"train",
"dev",
"test",
):
print("Processing {} subset...".format(subset)) print("Processing {} subset...".format(subset))
with open(Path(root_dir) / subset / "transcripts.txt") as fin: with open(Path(root_dir) / subset / "transcripts.txt") as fin:
subset_entries = [] subset_entries = []
@ -45,7 +49,7 @@ def main(root_dir):
/ audio_id_parts[1] / audio_id_parts[1]
/ "{}.opus".format(audio_id) / "{}.opus".format(audio_id)
) )
audio_filesize = os.path.getsize(audio_path) audio_duration = read_ogg_opus_duration(audio_path)
# TODO: support other languages # TODO: support other languages
transcript = ( transcript = (
transcript.strip() transcript.strip()
@ -76,7 +80,7 @@ def main(root_dir):
subset_entries.append( subset_entries.append(
( (
audio_path.relative_to(root_dir), audio_path.relative_to(root_dir),
audio_filesize, audio_duration,
transcript.strip(), transcript.strip(),
) )
) )