XiaoYuanZzz/IndexTTS_postrain

This is a brief introduction for model ckpts.

original models:

1. 1.bigvgan_discriminator.pth
2. 2.bigvgan_generator.pth
3. 3.bpe.model (tokenizer)
4. 4.config.yaml (original inference configuration)
5. 5.gpt.pth (backbone transformer)
6. 6.dvae.pth (use for getting training audio tokens)
7. 7.unigram_12000.vocab (tokenizer vocabulary)

fine-tuned models:

1. 1.Cantonese Model: a. configyue.yaml b. gpttunedyue.pth c. gptyue.pth (useless only for start stage in training)
2. 2.Janpanese Model: a. configja.yaml b. gpttunedja.pth c. gpttunedjav2.pth c. gpt_ja.pth (useless only for start stage in training)
3. 3.Malaya Model: a. configmalay.yaml b. gptmalayatuned.pth c. gptmalay.pth (useless only for start stage in training)