21 Commits

Author	SHA1	Message	Date
Kinan Martin	b167ac7b40	add utility file for creating subsets of mls english. must be fixed to make dev and test splits have matching sizes to reazonspeech	2025-07-28 17:52:36 +09:00
Kinan Martin	a8ecb16d47	use huggingface_hub library to download mls_english	2025-07-28 17:52:36 +09:00
Kinan Martin	f4b29870a0	switch mls_english clone from https to ssh	2025-07-28 17:52:36 +09:00
Kinan Martin	5417e0926b	restore version of mls_english compute_fbank_mls_english.py and prepare.sh from commit 547f5c5	2025-07-28 17:52:36 +09:00
Bailey Hirota	61e81bfc26	Revert "add fbank" ... This reverts commit ba603e0a0a514056ec6d32677053c41743a1a5dd.	2025-07-28 17:49:35 +09:00
Bailey Hirota	c83b115b49	add fbank	2025-07-28 17:49:35 +09:00
Kinan Martin	fa84782b21	optimize with num_jobs on save_audios	2025-07-28 17:49:35 +09:00
Kinan Martin	f2e01712de	fix stage 2 and 3	2025-07-28 17:49:35 +09:00
Kinan Martin	59519a41fa	fix validation manifest name	2025-07-28 17:49:35 +09:00
Kinan Martin	4ca8ee94f0	adjusted prepare.sh to only calculate fbank and manifest together; adjust datamodule to load from manifest files	2025-07-28 17:49:35 +09:00
Kinan Martin	d6e3c98e58	move compute_fbank_mls_english.py, add validate_manifest.py, add shared symlink to librispeech	2025-07-28 17:49:35 +09:00
Kinan Martin	68e3ceaaac	instead of on-the-fly features, precompute fbank and manifests in prepare.sh	2025-07-28 17:49:35 +09:00
Kinan Martin	ce44150e25	readme	2025-07-28 17:49:35 +09:00
Kinan Martin	a34d34a38e	pre-commit hooks	2025-07-28 17:49:35 +09:00
Kinan Martin	898525962c	separate transcript prep stage from bpe train stage	2025-07-28 17:49:35 +09:00
Kinan Martin	8c1c7100d3	symlink copied files to librispeech recipe dir	2025-07-28 17:49:35 +09:00
Kinan Martin	efe015d568	cleaned-up version of recipe	2025-07-28 17:49:35 +09:00
Kinan Martin	defc71bc6a	replace file	2025-07-28 17:49:35 +09:00
Kinan Martin	a1fc6420f9	change default path	2025-07-28 17:49:35 +09:00
Kinan Martin	ac0c0edddb	update prepare.sh, fix asr_datamodule.py	2025-07-28 17:49:35 +09:00
Kinan Martin	28f65458b3	WIP v0 MLS English recipe	2025-07-28 17:49:35 +09:00