models/Chroma - v1.0-HD

Chroma - v1.0-HD

9/3/2025

11:21:09 PM

Related Keywords & Tags

base model,base_model,checkpoint,chroma,chroma stable diffusion model,chroma1-base,chroma1-flash,chroma1-hd,chroma1-radiance,lodestone,other,v1.0-hd

Close-up digital painting of a dark muscular alien creature with a sleek shiny black exoskeleton, glowing yellow cracks on its elongated head, and a menacing toothy grin.

A dark gothic cityscape at sunset featuring a massive broken circular structure in the foreground, jagged spires and ruins silhouetted against a fiery orange-red sky with dramatic clouds, and rocky uneven terrain leading up to the structure.

Black and white photograph of a rusty metal swing hanging by two chains with raindrops, set against a blurred wooden fence background.

A surreal, haunting scene with dark, ominous creatures dancing on a glowing pathway leading to an ultrarealistic manor under a huge, colorful moon in a moonlit city.

Close-up of a menacing humanoid creature with gray decaying skin, sharp teeth, red glowing eyes, and pointed ears crouching in a forest with blurred green foliage background.

Black and white photograph of a weathered concrete wall with a large oval-shaped black hole in the center and a slightly ajar rusty metal door.

Black and white close-up photograph of a mechanical chrome spine focusing on central gears with shallow depth of field and blurred background.

Close-up photograph of a forest floor covered in autumn leaves with a large yellow maple leaf in the center, displaying water droplets, surrounded by smaller brown and red leaves with pine needles and twigs in the background.

Side view of a haunting figure with a distorted screaming face, head wrapped in cloth and draped in translucent fabric, surrounded by swirling smoke.

A grotesque pale segmented alien creature with twisted tentacle-like appendages lies on a wet metallic deck with water reflections, foggy sky, and a person’s leg in dark wet clothing visible.

Recommended Prompts

Extreme close-up photo of a tiger eye. large title "Chroma1" overlayed in the center of the image

Recommended Negative Prompts

low quality, ugly, unfinished, out of focus

Recommended Parameters

samplers

Heun

steps

cfg

resolution

512x512, 1024x1024

Tips

Chroma1-Base offers a solid foundation ideal for longer fine-tuning with high-res training at later epochs for faster convergence.

Chroma1-HD is recommended for quick fine-tuning or LoRA on high-resolution 1024x1024 images.

Chroma1-Flash applies delta weights to any Chroma model to speed up inference without GAN-based training; adjust strength accordingly.

Chroma1-Radiance is a work-in-progress pixel space model aimed at eliminating VAE compression artifacts.

Alternative quantized options include FP8 Scaled Quant and GGUF Quantized versions compatible with ComfyUI.

Creator Sponsors

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.