LayerDiffuse (transparent image layer diffusion)
Source: https://github.com/lllyasviel/layerdiffuse
Structured fields
{
"archived": false,
"revision": {
"ref": "main",
"pinned": false
},
"checked_at": "2026-10-17T00:00:00.000Z",
"usefulness": "Method and tools for generating transparent-background images and separable layers with diffusion models (latent transparency); a route to usable PNG-with-alpha stickers, overlays and layered illustration. Forge extension in a separate repo.",
"source_type": "repository",
"availability": "available",
"source_owner": "lllyasviel et al.",
"next_check_at": "2026-11-16T00:00:00.000Z",
"known_limitations": [
"No push since mid-2024; Forge extension flagged WIP."
],
"upstream_updated_at": "2024-06-16T08:05:00.000Z"
}Evidence claims
- [upstream_claim] README: LayerDiffuse: Transparent Image Layer Diffusion using Latent Transparency; points to a Forge extension and a Diffusers CLI repo. · locator: README source
- [sourced_fact] Repo metadata: Apache-2.0, not archived, last push 2024-06-16; the Forge extension lllyasviel/sd-forge-layerdiffuse (Apache-2.0, WIP, last push 2024-08-30) is also present. · locator: GitHub API fields source
- [sourced_fact] HF metadata: LayerDiffusion/layerdiffusion-v1 carries license:creativeml-openrail-m. · locator: HF API tags source
- [upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.
Rights
| component | license | hosting | redistribution | modification | training | Evidence and notes |
|---|---|---|---|---|---|---|
| code | Apache-2.0 | permitted | permitted | permitted | unknown | Apache-2.0. |
| model_weights | CreativeML-OpenRAIL-M | restricted | restricted | restricted | unknown | OpenRAIL-M use-based restrictions apply per HF tag; license text not read here. |