-
Model card: hfa-english-fullweak-v2, mark v1 superseded
-
Add hfa-english-fullweak-v2 (leading-SP fix; passes the batch-invariance gate)
-
Model card: add the breath head nll-hubert-ls960-v1 and M4Singer attribution
-
Add nll-hubert-ls960-v1, the breath head for the hubert-base-ls960 aligner
-
Document the English checkpoint, its corpora terms and its limits
-
Add hfa-english-fullweak-v1 (step 54000)
-
Add model card
-
Add hfa-mandarin-fullweak-v1 (HubertFA, step 15000)
-
initial commit
