Provenance/licensing clarification for controlnet-canny-sdxl-1.0

Hello, and thank you for maintaining Diffusers and the released ControlNet checkpoints.
We are documenting the provenance/licensing chain for the exact checkpoint:
diffusers/controlnet-canny-sdxl-1.0
Pinned revision:
eb115a19a10d14909256db740ed109532ab1483c
Governed SafeTensors artifact:
diffusion_pytorch_model.fp16.safetensors
SHA-256:
b2e7d3921058a442cc80430d1ec8847f42599c705e2451c95e77cf4dcf8d6c25
The pinned model card body describes this as a ControlNet trained on stabilityai/stable-diffusion-xl-base-1.0, and the repository metadata declares license: openrail++.
However, the pinned metadata also contains an older base_model: runwayml/stable-diffusion-v1-5 field, while repository history later changed tags/license metadata to SDXL/Open RAIL++ without removing that older field.
Could you please confirm the following factual points for this exact released checkpoint/revision:
Was controlnet-canny-sdxl-1.0 trained from / for stabilityai/stable-diffusion-xl-base-1.0, and is the base_model: runwayml/stable-diffusion-v1-5 field stale/incorrect metadata for this checkpoint?
Is the model-card declaration license: openrail++ intended to govern the actual released checkpoint weights in this repository, including the SafeTensors artifact above?
If available, which exact SDXL Base 1.0 revision/checkpoint was used for training this released ControlNet?
We are not asking for legal advice. We are trying to document the factual model lineage and the intended license declaration for the exact released artifact.
Thank you.

Start by reviewing the pinned revision eb115a19a10d14909256db740ed109532ab1483c, its model-card metadata, and the repository history for controlnet-canny-sdxl-1.0. Done means documenting whether the SDXL base_model field is stale, whether openrail++ applies to the named SafeTensors artifact, and which SDXL Base 1.0 revision was used if that information is available.

Scritto dal modello di indicizzazione a partire dal testo della issue.