Sd15 Inpainting FP16
Inpainting variant of the SD 1.5 unet, mirrored for the Sona Forge Android app.
Source du modèle
- Type
- Other
- Taille du fichier
- 1,60 GB
- Version
- vModel
- Mis à jour
- 9 oct. 2026
- Source
- Hugging Face
Description de la source
Stable Diffusion 1.5 Inpainting — ONNX FP16 (sona-forge mirror)
Inpainting variant of the SD 1.5 unet, mirrored for the Sona Forge Android app.
Sona Forge fetches every runtime binary from a single mirror under
huggingface.co/sona-forge so that
provider-side changes can't break end-user installs.
Provenance
unet/model.onnx is a byte-identical mirror of:
RanaLLC/stable-diffusion-v1-5-inpainting-onnx-fp16@a612b433500340de363966b9305a204e734822f6
which is an ONNX FP16 conversion of the Stable Diffusion 1.5 Inpainting
weights from runwayml/stable-diffusion-inpainting, released under
CreativeML Open RAIL-M.
The remaining inpainting-pipeline components (text_encoder, vae_decoder,
vae_encoder, tokenizer, scheduler) are reused from
sona-forge/sd15-base-fp16,
since the inpainting variant only diverges in the unet.
Sources
1 source
Sources
1 sourceVérifié 9 oct.
Artefacts du modèle
1 artefact
Artefacts du modèle
1 artefactExtraits de sources
2 extraits
Extraits de sources
2 extraits--- license: creativeml-openrail-m tags: - stable-diffusion - inpainting - onnx - sona-forge --- # Stable Diffusion 1.5 Inpainting — ONNX FP16 (sona-forge mirror) Inpainting variant of the SD 1.5 unet, mirrored for the [Sona Forge](https://github.com/virtaava/mobile-ai-image-app) Android app. Sona Forge fetches every runtime binary from a single mirror under [`huggingface.co/sona-forge`](https://huggingface.co/sona-forge) so that provider-side changes can't break end-user installs. ## Provenance `unet/model.onnx` is a byte-identical mirror of: - `RanaLLC/stable-diffusion-v1-5-inpainting-onnx-fp16` @ `a612b433500340de363966b9305a204e734822f6` which is an ONNX FP16 conversion of the Stable Diffusion 1.5 Inpainting weights from `runwayml/stable-diffusion-inpainting`, released under [CreativeML Open RAIL-M](https://huggingface.co/spaces/CompVis/stable-diffusion-license). The remaining inpainting-pipeline components (text_encoder, vae_decoder, vae_encoder, tokenizer, scheduler) are reused from [`sona-forge/sd15-base-fp16`](https://huggingface.co/sona-forge/sd15-base-fp16), since the inpainting variant only diverges in the unet.
Source context: 0 downloads · 0 likes · Repo sona-forge/sd15-inpainting-fp16
