TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Fooocus: OSS for image generation by ControlNet author

186 pointsby dvrpalmost 2 years ago

15 comments

erwannmillonalmost 2 years ago
&quot;Native refiner swap inside one single k-sampler. The advantage is that now the refiner model can reuse the base model&#x27;s momentum (or ODE&#x27;s history parameters) collected from k-sampling to achieve more coherent sampling. In Automatic1111&#x27;s high-res fix and ComfyUI&#x27;s node system, the base model and refiner use two independent k-samplers, which means the momentum is largely wasted, and the sampling continuity is broken. Fooocus uses its own advanced k-diffusion sampling that ensures seamless, native, and continuous swap in a refiner setup.&quot;<p>This is so interesting and seems obvious in retrospect, but super impressive! The code is simple too, going to hack around with this over the weekend :)
评论 #37097357 未加载
评论 #37097021 未加载
airgapstopgapalmost 2 years ago
&gt; Linux and Mac<p>&gt; Coming soon ...<p>Ah well. Hopefully it is soon. Also, on behalf of all Apple Silicon Mac users, would be nice if the author looked into implementing Metal FlashAttention [1].<p>1. <a href="https:&#x2F;&#x2F;github.com&#x2F;philipturner&#x2F;metal-flash-attention">https:&#x2F;&#x2F;github.com&#x2F;philipturner&#x2F;metal-flash-attention</a>
评论 #37098258 未加载
politelemonalmost 2 years ago
For those who don&#x27;t know, ControlNet is often used in conjunction with Stable Diffusion. It lets you add extra conditions to guide what is being generated. There are extensions for Automatic1111&#x27;s stable diffusion webui that can make use of ControlNet. Some examples I&#x27;ve seen are copying the pose of a person&#x2F;animal in an image and outputting a different person with the same pose (and extending to videos). Also taking line art drawings and filling it in with style.<p><a href="https:&#x2F;&#x2F;stable-diffusion-art.com&#x2F;controlnet&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;stable-diffusion-art.com&#x2F;controlnet&#x2F;</a>
评论 #37099278 未加载
kashunstvaalmost 2 years ago
&gt; Learned from Midjourney, the manual tweaking is not needed, and users only need to focus on the prompts and images<p>Except prompt-based tweaking doesn’t work very well in MJ; certainly not as well as manually-directed in-painting and out-painting. It’s virtually impossible in MJ to hold one part of the image constant while adding to&#x2F;modifying the remainder.
AbraKdabraalmost 2 years ago
Those commits are something else.
评论 #37098551 未加载
评论 #37097706 未加载
Hard_Spacealmost 2 years ago
Interesting, and I look forward to using it, but I wish the distribution had kept the folder-name conventions of AUTOMATIC1111, so that we could more easily have used symbolic links for folders of LoRAs and checkpoints etc. that we&#x27;d rather not duplicate.
评论 #37099174 未加载
评论 #37098626 未加载
评论 #37097596 未加载
GaggiXalmost 2 years ago
The names given to the commits are... peculiar.
评论 #37097374 未加载
评论 #37097241 未加载
评论 #37097356 未加载
yellow_postitalmost 2 years ago
definitely the smoothest install process and relatively snappy on my local windows machine that I&#x27;ve come across. I do hope to see some ControlNet integrations as that&#x27;s become a key part of my workflow for exploring new images.
评论 #37103539 未加载
captn3m0almost 2 years ago
Are there ways to run such apps with a remote GPU over network? I want to run the UI on my laptop, but use my homeserver GPU from the local network.<p>Anything better than X forwarding?
评论 #37097904 未加载
sorenjanalmost 2 years ago
Just like I expected, I get this error when trying to run it on my AMD GPU...<p>&quot;RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from <a href="http:&#x2F;&#x2F;www.nvidia.com&#x2F;Download&#x2F;index.aspx" rel="nofollow noreferrer">http:&#x2F;&#x2F;www.nvidia.com&#x2F;Download&#x2F;index.aspx</a>&quot;<p>Maybe it can get modified to use DirectML? Although it looks like it&#x27;s using PyTorch 2.0, and I think torch-directml only supports 1.13. Why is ML and GPGPU such a dependency mess?
bufferoverflowalmost 2 years ago
The sample image on the github page doesn&#x27;t look great. Major problems with the eyes, something both SD and MJ have solved, for the most part.
评论 #37101615 未加载
natchalmost 2 years ago
Great steps. I would still like to see something offline that can blend two disparate images into one generated scene, like artbreeder has.
评论 #37097635 未加载
r-k-joalmost 2 years ago
Here&#x27;s a live demo on HuggingFace <a href="https:&#x2F;&#x2F;huggingface.co&#x2F;spaces&#x2F;SpacesExamples&#x2F;Fooocus" rel="nofollow noreferrer">https:&#x2F;&#x2F;huggingface.co&#x2F;spaces&#x2F;SpacesExamples&#x2F;Fooocus</a>
burkayguralmost 2 years ago
Made a Discord bot with this. Check it out here <a href="http:&#x2F;&#x2F;fooocus.ai" rel="nofollow noreferrer">http:&#x2F;&#x2F;fooocus.ai</a>
brucethemoose2almost 2 years ago
I wonder if some of this can be ported to HF diffusers.<p>Lots of the changes just... make sense.