9 projects
mlx-teacache
TeaCache step-skipping for FLUX, Qwen-Image & Z-Image diffusion on Apple Silicon, in pure MLX
mlx-model-doctor
Validate MLX model repositories before loading them.
mlx-guard
External runtime safety supervision for MLX workloads on Apple Silicon
mlx-taef
Tiny AutoEncoders for diffusion on Apple Silicon — live previews + low-memory decode for FLUX, SD, and Qwen-Image.
mlx-dfloat
Run DFloat11 losslessly compressed BF16 checkpoints on Apple Silicon with MLX and a Metal decode kernel: image generation with FLUX.1, Z-Image, FLUX.2 Klein, Qwen-Image-2.1, ERNIE-Image and Krea 2 on a 32 GB Mac, with weights about 30% smaller
mlx-quant-fidelity
Measure quantization quality loss on Apple Silicon MLX — KL divergence, top-token flip rate and perplexity delta for KV-cache and weight quantization
mflux-taef
Live TAEF previews for mflux's command line, powered by mlx-taef.
mflux-teacache
TeaCache step skipping for mflux's command line, powered by mlx-teacache.
mlx-train-perf
Logit-free cross-entropy, sequence packing, Qwen 3.5 GatedDelta training and a RAM-fit planner for MLX LoRA fine-tuning