fp8 version of IllustriousXL v0.1

See a notebook on how to quantize.

All layers are quantized, so this may affect the generation quality.

Models

float8_e4m3fn is recommended.

Filename	Data type	Base weight
Illustrious-XL-v0.1.float8_e4m3fn.safetensors	float8_e4m3fn	Illustrious-XL-v0.1.safetensors
Illustrious-XL-v0.1.float8_e5m2.safetensors	float8_e5m2	Illustrious-XL-v0.1.safetensors
Illustrious-XL-v0.1-GUIDED.float8_e4m3fn.safetensors	float8_e4m3fn	Illustrious-XL-v0.1-GUIDED.safetensors
Illustrious-XL-v0.1-GUIDED.float8_e5m2.safetensors	float8_e5m2	Illustrious-XL-v0.1-GUIDED.safetensors

Tested with ComfyUI.

Number	Original(fp16)	float8_e4m3fn	float8_e5m2
1	👁 Image	👁 Image	👁 Image
2	👁 Image	👁 Image	👁 Image
3	👁 Image	👁 Image	👁 Image

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Base model

Finetuned

Quantized

(2)

this model