Full BF16 version:

chroma1_radiance_x16_calibrated.safetensors

Chroma1-Radiance (calibrated)

A lightly-calibrated checkpoint of Chroma1-Radiance by @LodestoneRock.

All credit for the underlying model goes to lodestones โ€” please support their work at:

sample 33 sample 34

What this is

The objective was:

  1. Improve overall visual quality โ€” sharper details
  2. Reduce leftover stochastic noise โ€” the high-frequency speckle that occasionally survives the late denoising steps in pixel-space generation.
  3. Reduce visibility of 16-px patch grid artifacts โ€” the regular checkerboard pattern from Radiance's patch tokenizer, especially in flat regions.

Files

File Size Format Use
chroma1_radiance_x16_calibrated.safetensors 19 GB BF16 Full precision, drop-in replacement for the base safetensors
chroma1_radiance_x16_calibrated_F16.gguf 19 GB F16 Full precision, GGUF
chroma1_radiance_x16_calibrated_Q8_0.gguf 10.1 GB GGUF Q8_0 Use with city96/ComfyUI-GGUF, ~half size, near-lossless
chroma1_radiance_x16_calibrated_Q5_1.gguf 7.14 GB GGUF Q5_1 Smaller GGUF for tighter VRAM (~8 GB cards with offload)

Pixel-space โ€” no VAE needed. Same workflow as vanilla Chroma1-Radiance.

Sample outputs

sample 1 sample 2 sample 3 sample 4 sample 5 sample 6 sample 7 sample 8 sample 9 sample 10 sample 11 sample 12 sample 13 sample 14 sample 15 sample 16 sample 17 sample 18 sample 19 sample 21 sample 21 sample 22 sample 23 sample 24 sample 25 sample 26 sample 27 sample 28 sample 29 sample 30 sample 31 sample 32

License

Inherits from Chroma1-Radiance โ€” Apache-2.0.

Downloads last month
13
GGUF
Model size
10B params
Architecture
flux
Hardware compatibility
Log In to add your hardware

5-bit

8-bit

16-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for wertz6969/chroma1_radiance_x16_calibrated

Quantized
(8)
this model