← ImageLoRA · All LoRA guides

Krea 2 LoRA training with OneTrainer: setup guide

Setting up OneTrainer to train a Krea 2 LoRA: support status, the settings known to work, and links to the docs. Status in OneTrainer: supported (Krea 2 16 GB preset benchmarked in OneTrainer PR #1680; not in README list).

About OneTrainer

GUI + CLI trainer with a broad model list; good VRAM use and speed, but more settings to configure. Sources: github.com/nerogar/OneTrainer, sanj.dev/post/onetrainer-vs-kohya-ss-vs-ai-toolkit/

OneTrainer-specific notes for Krea 2

• OneTrainer preview-2 branch (not master yet): LoRA on 8-bit quantized Krea 2 trains ~1.2x faster (16 GB preset); FLUX.2 16 GB preset ~1.3x; faster 8-bit math on RTX 50-series; DoRA not sped up

VRAM

ai-toolkit: ~20 GB at batch 1 with qfloat8; also trains on 12 GB with Krea Raw, low VRAM on, 50% transformer offload, 100% text-encoder offload, convrot 8, compile off, LoKr factor 8, 512 resolution only, and cached latents plus text embeddings. musubi: 12GB+ recommended for image training (general guidance); fp8 or --convrot_int8 halves weight VRAM

Settings that work for Krea 2 (any trainer)

• Natural-language captions, ~70-500 words; shorter captions lock features harder
• Diverse dataset (varied expressions/angles)
• Train at 1024px or higher
• Batch size >= 2 if VRAM allows
• Training at 512px can work for output up to ~2MP; add higher-res buckets gradually
• ai-toolkit downscales images larger than 1536px
• Automagic3 can give weaker LoRAs than v1/v2; pair it with LoKr factor 8 or 4 and sigmoid timesteps (that's my recipe). ai-toolkit's experimental automagic4 (Oct 23) keeps the LR from drifting to the floor
• EMA in ai-toolkit makes no noticeable difference; Automagic v3 already behaves a lot like EMA
• OneTrainer preview-2 branch (not master yet): LoRA on 8-bit quantized Krea 2 trains ~1.2x faster (16 GB preset); FLUX.2 16 GB preset ~1.3x; faster 8-bit math on RTX 50-series; DoRA not sped up
• Character captions: name, expression, pose, clothing, then scene
• Kaloscope style adapters (heathcliff01) for Krea 2 Raw and Turbo: style by artist name, weighted blends or a reference image; needs its own ComfyUI nodes
• ai-toolkit: point to local or alternative Qwen3-VL text encoder and Qwen-Image VAE files with text_encoder_path and vae_path under model_kwargs in the YAML
• Very slow steps (hundreds of seconds per step) mean the GPU is out of VRAM and swapping; lower the batch size or turn on low VRAM / offloading
• Pose control for Krea 2 Turbo: the Krea-2-PoseControl LoRA (MIT, 17-keypoint skeleton) with the ComfyUI-Krea2-Pose node pack; use release/krea2-pose-control-mix025.safetensors
• 12 GB ai-toolkit recipe (Krea Raw): low VRAM on, 50% transformer offload, 100% text-encoder offload, convrot 8 quantization, compile off, LoKr factor 8, 512 resolution only, cache latents and text embeddings. First steps can take many minutes; if the run hangs after the first step, try dropping the training adapter, a lower quant (Q6/Q4), or skipping text-encoder offload once embeddings are cached

General tips

• 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

Compare with other trainers

• ai-toolkit vs OneTrainer for Krea 2
• OneTrainer vs musubi-tuner for Krea 2
• OneTrainer vs SimpleTuner for Krea 2
• Krea 2 LoRA with ai-toolkit
• Krea 2 LoRA with musubi-tuner
• Krea 2 LoRA with SimpleTuner
• Krea 2 (overview)

Where to rent a 16 GB+ GPU

Cheapest 16 GB+ cards (2026-10-20 08:06 UTC) from RunPod and Vast.ai. Prices move constantly; see the live price table.

Sources

• blog.hinablue.me/krea2-lora-trainer/
• www.reddit.com/r/StableDiffusion/comments/1vrkd5x/which_platform_shoul
• sanj.dev/post/onetrainer-vs-kohya-ss-vs-ai-toolkit/
• github.com/ostris/ai-toolkit#readme
• github.com/kohya-ss/musubi-tuner#readme
• github.com/kohya-ss/musubi-tuner/blob/main/docs/krea2.md
• github.com/bghira/SimpleTuner#readme
• github.com/Nerogar/OneTrainer/pull/1680
• huggingface.co/heathcliff01/Kaloscope-style-adapter
• github.com/ostris/ai-toolkit/commit/534d462ff622e334e1a65b42884d96d983
• github.com/ostris/ai-toolkit/blob/main/extensions_built_in/diffusion_m
• huggingface.co/adhit-420/Krea-2-PoseControl-LoRA-checkpoints
• github.com/nerogar/OneTrainer

FAQ

Does OneTrainer support Krea 2?

Yes: supported (Krea 2 16 GB preset benchmarked in OneTrainer PR #1680; not in README list) (see sources below).

How much VRAM do I need for a Krea 2 LoRA?

ai-toolkit: ~20 GB at batch 1 with qfloat8; also trains on 12 GB with Krea Raw, low VRAM on, 50% transformer offload, 100% text-encoder offload, convrot 8, compile off, LoKr factor 8, 512 resolution only, and cached latents plus text embeddings. musubi: 12GB+ recommended for image training (general guidance); fp8 or --convrot_int8 halves weight VRAM

Data updated 2026-10-20. Disclosure: RunPod and Vast.ai links are referral links; ImageLoRA may earn a commission at no cost to you. unknown = no data yet. · Raw data (JSON) · llms.txt