Nodes/CRT-Nodes/DepthAnything3 (CRT)
ComfyUI Node

DepthAnything3 (CRT)

A ComfyUI node in CRT/DepthAnything3 with 12 inputs and 1 output.

By PGCRT·Created 2 years ago·Updated about 11 hours ago· 132
DepthAnything3 (CRT)
  • image
  • image
model_namedepth_anything_3_small.safetensors
keep_model_loadedtrue
weight_dtypedefault
megapixels0.50
modemono
outputdepth
normalizationv2_style
apply_sky_clipfalse
coloredfalse
ref_view_strategysaddle_balanced
pose_methodcam_dec
CategoryCRT/DepthAnything3

Inputs (12)

NameTypeDefaultDescription
model_nameCOMBOdepth_anything_3_small.safetensors4 options: depth_anything_3_small.safetensors, depth_anything_3_base.safetensors, depth_anything_3_mono_large.safetensors, depth_anything_3_metric_large.safetensors
keep_model_loadedBOOLEANtrue
weight_dtypeCOMBOdefault4 options: default, fp16, bf16, fp32
imageIMAGE
megapixelsFLOAT0.500.05–8
modeCOMBOmono2 options: mono, multiview
outputCOMBOdepth4 options: depth, depth_colored, sky_mask, confidence
normalizationCOMBOv2_style3 options: v2_style, min_max, raw
apply_sky_clipoptBOOLEANfalse
coloredoptBOOLEANfalse
ref_view_strategyoptCOMBOsaddle_balanced4 options: saddle_balanced, saddle_sim_range, first, middle
pose_methodoptCOMBOcam_dec2 options: cam_dec, ray_pose

Outputs (1)

NameTypeDescription
imageIMAGE