# Help implementing Tiled Diffusion and Tiled VAE with Diffusers **URL:** \ **Category:** Beginners\ **Created:** [April 14, 2025, 9:45am UTC](https://discuss.huggingface.co/t/help-implementing-tiled-diffusion-and-tiled-vae-with-diffusers/150354 "2025-04-14T09:45:29Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![OxiiLB](https://avatars.discourse-cdn.com/v4/letter/o/278dde/32.png) [@OxiiLB](https://discuss.huggingface.co/u/OxiiLB)\ **Post date:** [April 14, 2025, 9:45am UTC](https://discuss.huggingface.co/t/help-implementing-tiled-diffusion-and-tiled-vae-with-diffusers/150354/1 "2025-04-14T09:45:29Z")
Hi everyone, I’m working on a Python script using the 🤗 Diffusers library. My goal is to take an input image and return an upscaled, more detailed version of it, essentially magnified and enhanced. I’m using the Juggernaut Reborn checkpoint and would like to implement Tiled Diffusion and Tiled VAE to improve performance and quality on high-resolution images. Additionally, I’m using two LoRAs: more\_details and SDXLrender\_v2.0. The problem is: I’m having trouble finding up-to-date or clear documentation on how to integrate Tiled Diffusion and Tiled VAE with Diffusers, especially in an image2image pipeline. Has anyone successfully implemented this setup? Any example code, resources, or guidance would be greatly appreciated! Thanks in advance! ---
**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](https://discuss.huggingface.co/u/John6666)\ **Post date:** [April 14, 2025, 10:01am UTC](https://discuss.huggingface.co/t/help-implementing-tiled-diffusion-and-tiled-vae-with-diffusers/150354/2 "2025-04-14T10:01:58Z")
It’s going to be implemented soon, but it’s not there yet… > > > \*\*Is your feature request related to a problem? Please describe.\*\* > Currently, d…iffusers library lacks advanced tiled upscaling capabilities that are available in other Stable Diffusion implementations. While the library supports basic img2img and upscaling, there's no built-in solution for handling large images through intelligent tiling and seam fixing. This makes it difficult to process high-resolution images while maintaining quality and managing memory efficiently. > > \*\*Describe the solution you'd like\*\* > Implement Ultimate SD Upscale functionality (similar to Automatic1111's WebUI extension) as a pipeline in diffusers. Key features should include: > > 1. Progressive upscaling with intelligent scale factor determination > 2. Multiple tiling modes: > - Linear processing > - Chess pattern processing > 3. Advanced seam fixing options: > - Band pass mode > - Half tile offset > - Half tile with intersections > 4. Configurable parameters: > - Tile sizes > - Padding > - Mask blur > - Denoise strength for seam fixing > > The implementation should integrate smoothly with existing diffusers pipelines and maintain the library's user-friendly API style. > > \*\*Describe alternatives you've considered\*\* > \- Using basic tiling without seam fixing (leads to visible artifacts) > \- Running multiple separate upscale passes (inefficient and lower quality) > \- Implementing as a separate package (loses benefits of diffusers' optimization and integration) > \- Using other libraries like PIL or cv2 for tiling (lacks SD-specific optimizations) > > \*\*Additional context\*\* > \- Reference implementation: https://github.com/Coyote-A/ultimate-upscale-for-automatic1111 > \- This feature would be particularly valuable for: > - Professional image upscaling > - Batch processing of large images > - Creating high-resolution outputs while managing VRAM > - Maintaining image quality in tiled processing > \- Could potentially be implemented as either a standalone pipeline or an enhancement to existing img2img pipelines > \- Would complement existing super-resolution models in the library > **[GitHub - gokayfem/Tile-Upscaler: Image Upscaler with Tile Controlnet Fully...](https://github.com/gokayfem/Tile-Upscaler)** > > Image Upscaler with Tile Controlnet Fully Integrated in Huggingface Diffusers ---
**Author:** ![OxiiLB](https://avatars.discourse-cdn.com/v4/letter/o/278dde/32.png) [@OxiiLB](https://discuss.huggingface.co/u/OxiiLB)\ **Post date:** [April 14, 2025, 10:10am UTC](https://discuss.huggingface.co/t/help-implementing-tiled-diffusion-and-tiled-vae-with-diffusers/150354/3 "2025-04-14T10:10:14Z")
Thank you for the update. I’m working with high-resolution images, and I was wondering if you have any suggestions on how to proceed without Tiled Diffusion for now? ---
**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](https://discuss.huggingface.co/u/John6666)\ **Post date:** [April 14, 2025, 10:28am UTC](https://discuss.huggingface.co/t/help-implementing-tiled-diffusion-and-tiled-vae-with-diffusers/150354/4 "2025-04-14T10:28:50Z")
If you want to complete it with Diffusers, this is probably the only way. There are quite a few AI models for upscaling, so I think there are also ways to use them separately. > **[Latent upscaler](https://huggingface.co/docs/diffusers/main/en/api/pipelines/stable_diffusion/latent_upscale)** > > We’re on a journey to advance and democratize artificial intelligence through open source and open science. > **[Spaces - Hugging Face](https://huggingface.co/spaces?q=upscale)** > > Discover amazing ML apps made by the community