Pinned Loading

1. Polyvox-Studio Polyvox-Studio Public

   🎙️ Train text-to-speech voices from your own recordings and use them, all on your own computer. Turns raw audio into a clean dataset, fine-tunes VoxCPM2, dots.tts, GPT-SoVITS or Fish Audio S2-Pro, …

   Python

2. LoRA-Forge LoRA-Forge Public

   Web UI for training LoRAs on Anima, Krea 2, MOSS-Image, SDXL, FLUX, SD3.5 and SD1.5. Tuned for speed: per-block torch.compile, flash-attn, fp16 accumulation and a batched ProdigyPlus step (1.5-2x f…

   TypeScript 4