AI BriefingRelayVSR: RelayVSR repackaged for ComfyUI, 2x·4x video upscaling
Comfy-Org/RelayVSR
About the project
A file set repackaged to allow direct use of the RelayVSR model in the ComfyUI environment. The model files from the original repository have been organized to match ComfyUI's directory structure.
It provides two model files supporting 2x and 4x resolution upscaling, along with their corresponding VAE files. It includes int8 quantized versions and fp32 VAEs, intended for use in ComfyUI workflows to increase video resolution.
Distributed under the Apache-2.0 license, the files can be placed in ComfyUI's diffusion_models and vae folders to be called immediately in workflows. The goal is to integrate into existing video upscaling pipelines without any code modifications.
Comfy-Org/RelayVSR
The original page has no description.
unconditional-image-generation
This introduction was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report errors, attribution issues, or removal requests via Contact.