LTX-2 Phr00t NSFW v62 (MLX Version)
This repository contains Mac-native MLX quantizations of Phr00t's LTX-2 NSFW v62 merge. Optimized for Apple Silicon (M1/M2/M3).
Which version should I use?
| Version | Folder | RAM Required | Best For |
|---|---|---|---|
| 4-Bit | 4bit | 16GB+ | Standard MacBook Pros (M1/M2/M3 Pro) |
| 8-Bit | 8bit | 32GB+ | M3 Max / Ultra users seeking maximum quality |
How to use in James App
- Download the
model.safetensorsfrom the folder matching your RAM. - Rename it to
diffusion_pytorch_model.safetensors. - Replace your existing file in the app's snapshot folder.