Explore how a color-depth LoRA changes lighting and palette in a MOSS-Image SFT text-to-image workflow. The LoRA is connected in the output path, so this JSON represents an active color-control experiment.

Main features:
- MOSS-Image SFT INT8 ConvRot model, Qwen3.5 9B BF16 text encoder, and FLUX.2 VAE.
- moss_image_color_depth_control.safetensors connected at strength 0.8.
- One output-reachable Euler sampler: 50 steps, CFG 3, simple scheduler, fixed seed 14001.
- Example neon rain-street prompt with magenta and cyan reflections, plus a separate negative prompt.

Suggested workflow:
1. Install the SFT model, text encoder, VAE, and color-depth LoRA, then load the JSON.
2. Generate with the included prompt and LoRA strength 0.8 to establish a baseline.
3. Change subject, lighting words, and LoRA strength to compare color depth and contrast in SaveImage outputs.

Required model filenames: moss_image_sft_int8_convrot.safetensors; qwen3.5_9b_bf16.safetensors; flux2-vae.safetensors; moss_image_color_depth_control.safetensors. The cover is a related example image; the JSON's current prompt is the neon rain street scene.

RunningHub AI workflow: https://www.runninghub.ai/zh-cn/post/2113906855339978754?inviteCode=rh-v1111
RunningHub signup: https://www.runninghub.ai/?inviteCode=rh-v1111
Bilibili video: https://www.bilibili.com/video/BV1ZHHP6FEVZ/
Related model files: https://pan.quark.cn/s/20e19307e8ef