All-in-one TRELLIS2 conditioning with built-in model config.
Combines model configuration + DinoV3 feature extraction in one node.
Model config is embedded in conditioning - no separate config needed!
SINGLE WORKFLOW:
One node → connect conditioning to both Image to Shape and Shape to Textured Mesh.
DUAL WORKFLOW (recommended for quality):
1. First node: resolution=512 → fast shape generation
- Connect conditioning → Image to Shape
2. Second node: resolution=1536_cascade, different image if desired
- Connect conditioning → Shape to Textured Mesh (texture_conditioning)
This gives fast shape (~15s) with high-detail texture voxelgrid.
Resolution modes:
- 512: Fast (~15s), lower detail - good for shape
- 1024_cascade: Balanced (~30s) - good default
- 1536_cascade: Highest detail (~45s) - best for texture
By BizaNator·Created 7 months ago·Updated 16 days ago· 14
BD TRELLIS.2 Get Conditioning
image
mask
conditioning
preprocessed_image
◄resolution1024_cascade►
◄attn_backendflash_attn►
◄vram_modekeep_loaded►
◄include_1024true►
◄background_colorgray►
Category🧠BrainDead/TRELLIS2
Inputs (7)
Name
Type
Default
Description
image
IMAGE
—
mask
MASK
—
resolution
COMBO
1024_cascade
Model resolution. Use 512 for fast shape, 1536_cascade for detailed texture.
attn_backend
COMBO
flash_attn
Attention implementation
vram_mode
COMBO
keep_loaded
VRAM usage strategy
include_1024
BOOLEAN
true
Extract 1024px features (needed for cascade modes)