speecht5_tts-fsc

microsoft/speecht5_tts finetuned for Filipino (Tagalog/Taglish) on sapinsapin/filipinospeechcorpus.

Trained for 1000 steps on 1867 read-speech clips (batch 4x8, lr 1e-05, fp32 + gradient checkpointing). Synthesized listen-test samples are in samples/ (speechbrain x-vector speaker conditioning + microsoft/speecht5_hifigan vocoder).

metric	value
eval_loss	0.4432

Trained with finetune_tts.py from the halohalo pipeline; the dataset adapter normalizes each corpus to (audio@16k, text, speaker_id) so corpora are swappable with a --dataset flag.