This is the official model page for
Chroma1-HD GGUF Quants
Workflow is in the training data .zip file
The following is a copy of original model page.
Hey everyone,
A while back, I posted about Chroma, my work-in-progress, open-source foundational model. I got a ton of great feedback, and I'm excited to announce that the base model training is finally complete, and the whole family of models is now ready for you to use!
A quick refresher on the promise here: these are true base models.
I haven't done any aesthetic tuning or used post-training stuff like DPO. They are raw, powerful, and designed to be the perfect, neutral starting point for you to fine-tune. We did the heavy lifting so you don't have to.
And by heavy lifting, I mean about 105,000 H100 hours of compute. All that GPU time went into packing these models with a massive data distribution, which should make fine-tuning on top of them a breeze.
As promised, everything is fully Apache 2.0 licensed—no gatekeeping.
TL;DR:
Release branch:
Chroma1-Base: This is the core 512x512 model. It's a solid, all-around foundation for pretty much any creative project. You might want to use this one if you’re planning to fine-tune it for longer and then only train high res at the end of the epochs to make it converge faster.
Chroma1-HD: This is the high-res fine-tune of the Chroma1-Base at a 1024x1024 resolution. If you're looking to do a quick fine-tune or LoRA for high-res, this is your starting point.
Research Branch:
Chroma1-Flash: A fine-tuned version of the Chroma1-Base I made to find the best way to make these flow matching models faster. This is technically an experimental result to figure out how to train a fast model without utilizing any GAN-based training. The delta weights can be applied to any Chroma version to make it faster (just make sure to adjust the strength).
Chroma1-Radiance [WIP]: A radical tuned version of the Chroma1-Base where the model is now a pixel space model which technically should not suffer from the VAE compression artifacts.
Quantization options
Alternative option: FP8 Scaled Quant (Format used by ComfyUI with possible inference speed increase)
Alternative option: GGUF Quantized (You will need to install ComfyUI-GGUF custom node)
Special Thanks
A massive thank you to the supporters who make this project possible.
Anonymous donor whose incredible generosity funded the pretraining run and data collections. Your support has been transformative for open-source AI.
Fictional.ai for their fantastic support and for helping push the boundaries of open-source AI.
Support this project!
https://ko-fi.com/lodestonerock/
BTC address: bc1qahn97gm03csxeqs7f4avdwecahdj4mcp9dytnj
ETH address: 0x679C0C419E949d8f3515a255cE675A1c4D92A3d7
my discord: discord.gg/SQVcWVbqKx
Description
This fp8 scaled model is only usable in ComfyUI and in order to use it as intended, you need to install ComfyUI_Hybrid-Scaled_fp8-Loader
This will only worked with pruned flash-heun LoRAs
FAQ
Comments (8)
With so many models, if I only intend to use them without training any LoRA, which one is the best choice?
@573410705748 I've read about Flux, and logic says that it should be correct, that fp16 will be better then fp8, even quantized. And as you can see people either use full model, or Q8_0.
It was on S1LV3RC01N/silver's civitai account :/
What was?
What's the difference between this FP8 and the FP8 by Clybius on HF?
Also, what's the difference between the 3 FP8 versions here, aside of the use of LORAs?
I don't understand why my results look so absolutely horrendous, I do not understand, I just can't get it to work.
ComfyUI_Hybrid-Scaled_fp8-Loader
Error 404
im confused is this a gguf file or diffusion checkpoint?