Model description
These are Chunte/huggy-style-v6-lora DreamBooth LoRA weights for black-forest-labs/FLUX.1-dev.
The weights were trained using DreamBooth with the Flux diffusers trainer.
Was LoRA for the text encoder enabled? False.
Trigger words
You should use hggycrcl to trigger the image generation.
Featured example — Spock Huggy (Project ProtoStar)
A Starfleet science-officer Huggy giving the Vulcan salute, generated with this LoRA for the Project ProtoStar deck.
Prompt
hggycrcl, circular body, floating hands, orange outlined, darker orange cel shade on body, flat 2D, black dot eyes, red open mouth, happy open mouth, round ball-shaped body, only two floating hands, no arms, no torso, no shoulders, black bowl cut hair with straight bangs on top of the round head, small pointed vulcan ears, angled eyebrows, lower half of the round body colored as a blue starfleet uniform with a gold delta arrowhead badge, one floating hand raised palm forward in the vulcan salute, fingers parted two-and-two with a gap between the middle and ring finger, live long and prosper
Inference spec
- Base:
black-forest-labs/FLUX.1-dev · LoRA scale 1.0
- 768×768 · 28 steps · guidance 3.5
- Seed: 99
Prompt-engineering notes (what made it work)
- Keep Huggy anatomy. Lead with
round ball-shaped body, only two floating hands, no arms, no torso, no shoulders. Without this, adding hair + a costume drifts the model into a human chibi with a torso.
- Costume as body color. Color the lower half of the round body as the uniform; avoid "uniform top / shirt / wearing", which also induces a torso.
- Vulcan salute. Phrase it as
fingers parted two-and-two with a gap between the middle and ring finger + live long and prosper. Avoid "split into a V" — the model renders a peace sign instead.
- Spock features.
black bowl cut hair with straight bangs, small pointed vulcan ears, angled eyebrows.
- Reliable seeds for clean hands: 99, 555, 1701, 314, 2024. Hands are V6's weak spot — generate several seeds and pick.
Featured example — Red Ops Huggy (Project ProtoStar)
A Starfleet operations-division Huggy in a red uniform giving an enthusiastic thumbs up, generated with this LoRA for the Project ProtoStar deck.
Prompt
hggycrcl, circular body, floating hands, orange outlined, darker orange cel shade on body, flat 2D, black dot eyes, red open mouth, happy open mouth, round ball-shaped body, only two floating hands, no arms, no torso, no shoulders, lower half of the round body colored as a red operations starfleet uniform with a gold delta arrowhead badge, both floating hands giving a thumbs up
Inference spec
- Base:
black-forest-labs/FLUX.1-dev · LoRA scale 1.0
- 768×768 · 28 steps · guidance 3.5
- Seed: 1701
Prompt-engineering notes (what made it work)
- Same anatomy anchor as Spock.
round ball-shaped body, only two floating hands, no arms, no torso, no shoulders keeps it a floating blob — essential once a uniform is involved.
- Uniform as body color.
lower half of the round body colored as a red operations starfleet uniform with a gold delta arrowhead badge paints the costume onto the body without inducing a torso or sleeves.
- Gesture.
both floating hands giving a thumbs up reads cleanly; gestures involving both hands at once are more reliable than complex single-hand finger poses.
- Crew tip. Swap the uniform color to assemble a full crew —
gold command / blue starfleet / red operations — keeping everything else identical.
Download model
Download the *.safetensors LoRA in the Files & versions tab.
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16).to('cuda')
pipeline.load_lora_weights('Chunte/huggy-style-v6-lora', weight_name='pytorch_lora_weights.safetensors')
image = pipeline('hggycrcl').images[0]
For more details, including weighting, merging and fusing LoRAs, check the documentation on loading LoRAs in diffusers
License
Please adhere to the licensing terms as described here.
Intended uses & limitations
How to use
Limitations and bias
[TODO: provide examples of latent issues and potential remediations]
Training details
[TODO: describe the data used to train the model]
