How to train a FLUX.2 LoRA (dev)

Which trainers support FLUX.2 dev (32B), how much VRAM you need, and settings that work. Every row links to its source.

Trainer support

80 GB recommended; ai-toolkit ships train_lora_flux_24gb.yaml

Start from ai-toolkit train_lora_flux_24gb.yaml
OneTrainer preview-2 branch (not master yet): LoRA on 8-bit quantized FLUX.2 trains ~1.3x faster with the 16 GB preset; DoRA not sped up
diffusers v0.42.0 (Oct 23, PR #14163) loads LoKr adapters for FLUX.2 and FLUX.2 klein
ComfyUI master (PR #17196) runs the FLUX.2 VAE in fp16: faster, closer to fp32, and no tiled VAE node needed at high resolutions. Your existing VAE file works after updating; don't use bf16 FLUX.2 VAE files

Gated Hugging Face models: click 'Agree and access' on each model page, or you'll get 403 errors
Steps scale with dataset size (steps = epochs x images at batch 1)
A safe start: LoKr with auto rank (~64), batch/grad accumulation 4
Leave out of your captions whatever you want burned into the LoRA; caption what should stay changeable
ai-toolkit dataset folder weights let you train the same images with long, medium and short captions in separate folders
ai-toolkit supports validation loss, which shows when a run stops improving
In ai-toolkit, Cache text embeddings works for basically every model and frees the text encoder from VRAM and RAM
Prefer int8 convrot over qfloat8 quantization in ai-toolkit: better quality at the same memory
Gradient accumulation and batch size don't scale the learning rate in ai-toolkit; a bigger batch needs about 1.5-2.5x the steps

Weekly newsletter: new models, trainer releases, cheapest GPUs. Subscribe

Data updated 2026-10-20. RunPod and Vast.ai links are referral links. unknown = no data yet. · Raw data (JSON) · llms.txt