Apply for a GPU community grant: Academic project
#2
by speedmaker - opened
We built a public Hugging Face demo for IndexTTS2: English, Female Voice (Singaporean Accent).
It uses a fine-tuned checkpoint trained on an English female voice from IMDA’s open-source National Speech Corpus (NSC). Users only need to enter text; the demo uses a fixed built-in reference voice.
Our goal is to make Singaporean-accent TTS easy to test for educators, students, and developers.
Right now, the demo is too slow on limited hardware. A Community GPU Grant would let us provide faster, more reliable inference for more users, while keeping the model and documentation public and reproducible.