cd /news/generative-ai/truerelight-directional-physical-rel… · home › topics › generative-ai › article
[ARTICLE · art-170598] src=discuss.huggingface.co ↗ pub= topic=generative-ai verified=true sentiment=↑ positive

TrueRelight: Directional & Physical Relighting LoRA for FLUX.3 Image Klein (16-step distilled, Open Weights + 280-Pair Dataset + Interactive Space)

A developer publishing as pandacooming released TrueRelight, an open-source directional and physical relighting LoRA for Black Forest Labs' FLUX.3 Image Klein, trained on the flux-3-image-klein-base model and evaluated on the 16-step distilled flux-3-image-klein. The LoRA was trained on 280 pixel-aligned <ref_image_0> pairs at rank=32 and 1024px, and the author reports that popping the prompt_upsampler block during 16-step inference lets TrueRelight reverse light direction instead of producing the flat color wash seen with prompt-only relighting. The release includes a 160.7 MB pytorch_lora_weights.safetensors file, a 280-pair Hugging Face dataset, a demo notebook, and ComfyUI compatibility at strength 0.8–1.0 with 16 steps, CFG 1.0, and euler/simple samplers.

read1 min views4 publishedNov 16, 2026

Hi everyone! I just open-sourced TrueRelight, a physical & directional relighting LoRA trained on black-forest-labs/flux-3-image-klein-base and evaluated on the 16-step distilled black-forest-labs/flux-3-image-klein.

Plain prompt-only relighting on distilled FLUX.3 Image Klein often behaves like a flat color wash (e.g., tinting existing shadows orange/yellow instead of reversing light direction). By training on 280 pixel-aligned <ref_image_0> pairs (rank=32, 1024px) and popping prompt_upsampler during 16-step inference, TrueRelight can:

pytorch_lora_weights.safetensors, 160.7 MB): https://huggingface.co/datasets/pandacooming/FLUX-3-Image-Klein-TrueRelight-Dataset``https://huggingface.co/pandacooming/FLUX-3-Image-Klein-TrueRelight-LoRA/blob/main/FLUX_3_Klein_TrueRelight_Demo.ipynb``diffusers) Pop prompt_upsampler before running the 16-step pipeline so the <ref_image_0> grammar reaches the LoRA cleanly:

blocks = ModularPipeline.from_pretrained("black-forest-labs/flux-3-image-klein").blocks
blocks.sub_blocks.pop("prompt_upsampler", None)
pipe = blocks.init_pipeline("black-forest-labs/flux-3-image-klein")

(Also compatible with flux-3-image-klein-comfy at strength 0.8–1.0, 16 steps, CFG 1.0, euler/ simple, skipping the prompt upsampler.)

Would love feedback from anyone testing it on portraits, tricky shadows, or ComfyUI workflows!

── more in #generative-ai 4 stories · sorted by recency
── more on @truerelight 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/truerelight-directio…] indexed:0 read:1min 2026-11-16 · —