A big thank you to RunDiffusion! They sponsored the compute for Juggernaut Moss, the model this LoRA was made for. Without them none of this would exist.
This is my own DMD2 turbo LoRA for Juggernaut Moss. I didn't want to use one of the existing speed LoRAs. I wanted one made specifically for Juggernaut Moss, so I trained it myself. With the LoRA, Juggernaut Moss needs 6 steps at CFG 1 instead of 30 steps at CFG 3. On my RTX A4500 a 1344x1792 image takes about 11 seconds instead of more than a minute.
This is a Beta. The results are close to the full model, but the Turbo has its own character: it picks safer compositions and gives less variety between seeds. I trained it on the int8 version, so that is what it works best with. The next version will be trained on BF16 and I'm already working on improvements. If you find weak spots, please tell me.
💼 I'm looking for a job! I'm currently open to new opportunities, either a full-time position or long-term paid contract work in model training, finetuning and image generation. If you like what I build, reach out on LinkedIn or by email to KandooAI@web.de.
| File | Size | Use |
|---|---|---|
JuggernautMossDMD2Lora4-8Steps.safetensors |
297 MB | the LoRA (rank 64), for ComfyUI models/loras/ |
workflows/Juggernaut_Moss_Turbo_LoRA.json |
ComfyUI workflow, drag it into ComfyUI | |
workflows/Juggernaut_Moss_Turbo_LoRA_API.json |
the same workflow in API format | |
images/showcase/PROMPTS.txt |
the prompts and seeds of all showcase images |
If you don't want to load a LoRA, there will also be Juggernaut Moss Turbo, a checkpoint with this LoRA already baked in, in its own repo.
juggernaut_moss_v1_int8_convrot.safetensors + this LoRA at strength 1.0 (0.9 works about the same).qwen3.5_9b_int8_convrot.safetensors (or bf16 / w4a8_gptq) in models/text_encoders/ with Load CLIP, type moss_image, and flux2-vae_small.safetensors in models/vae/.The LoRA also works on plain MOSS-Image Base, but it was made for Juggernaut Moss and that is where it looks best.
Juggernaut Moss was trained on detailed natural-language captions, and the Turbo likes the same kind of prompt. The structure that works best follows the training captions:
Text in the image: write every word you want to see in quotes, for example a neon sign that reads "OPEN ALL NIGHT". And this matters even more with the Turbo: if there is a place where text would naturally appear (a sign, a label, a jersey, a book cover), prompt that text too. Otherwise the model fills it with gibberish. Telling it "no other text" does not work.
Example:
A cinematic film still of a street musician in his forties playing the saxophone alone under a stone
bridge at night, eyes closed, completely lost in the music. He wears a long dark coat and a knitted grey
hat, and an open saxophone case at his feet holds a few coins and a handwritten cardboard sign reading
"THANK YOU". The shot is a medium shot taken from a low angle in three-quarter profile. The lighting is
low-key, with warm orange light from a street lamp on one side of his face and cool blue light from the
river on the other. The image has a fine film grain and soft halation around the light sources. Shot
with a Sony ILCE-7M3, 35mm lens at f/1.4.
MOSS-Image can produce nudity, and neither Juggernaut Moss nor this LoRA removes that ability. That is why this repo is tagged not-for-all-audiences. Use it responsibly and within the law.
All images were made with Juggernaut Moss int8 + this LoRA at strength 1.0, 6 steps, CFG 1, euler_ancestral / simple, one seed per prompt, no upscaling and no inpainting. I picked them from several hundred renders. The prompts and seeds are in images/showcase/PROMPTS.txt.
Apache License 2.0, the same as Juggernaut Moss and MOSS-Image. You can use the LoRA and its outputs commercially.
KandooAI, here on Hugging Face: huggingface.co/Kandoo
Job offers and collaborations: LinkedIn or KandooAI@web.de
Kandoo/Juggernaut-Moss