toxicdog/MOSS-Image-SFT-Turbo-INT8-ConvRot

MOSS-Image-SFT-Turbo INT8 ConvRot

Experimental model files for ComfyUI and others, mainly created for testing in Radiant Canvas for fast macOS inference (https://apps.apple.com/us/app/radiant-canvas-ai-image-gen/id6802973075).

A fast 8-step variant of OpenMOSS-Team/MOSS-Image-SFT.

Use CFG 1 and 8 steps. No negative prompt is used.

File

• diffusion_models/moss_image_sft_turbo_int8_convrot.safetensors: the diffusion transformer as a single file in ComfyUI's native INT8 ConvRot format.

ComfyUI

MOSS-Image is supported natively in ComfyUI master since October 16, 2026. No custom nodes are needed. Use a ComfyUI build that includes MOSS-Image support and INT8 ConvRot.

Place the files in the following folders:

ComfyUI/models/diffusion_models/moss_image_sft_turbo_int8_convrot.safetensors
ComfyUI/models/text_encoders/qwen3.5_9b_int8_convrot.safetensors
ComfyUI/models/vae/flux2-vae.safetensors

The text encoder and VAE are the same as MOSS-Image-SFT's. Get them from Comfy-Org/MOSS-Image (qwen3.5_9b_w4a8_gptq.safetensors also works as the text encoder).

Start from the standard MOSS-Image workflow. Load this file with the diffusion model loader, load the text encoder with type moss_image, and set the sampler to:

Setting	Value
Steps	8
CFG	1
Sampler	euler
Scheduler	simple
Shift	5 (the model default)

The model uses the standard flow-matching sampler. No custom scheduler is needed.

Precision

• The 252 attention and MLP block weights are INT8 ConvRot (per-channel scale, ConvRot group size 256), quantized from FP32 weights.
• rope.bands is FP32. All other tensors are BF16.

This is an experimental release. Check output quality for your own prompts before relying on it.

License

Released under the Apache License 2.0. This is a modified version of MOSS-Image-SFT by OpenMOSS, which is also licensed under the Apache License 2.0. A copy of the license is included in this repository.