Nodes/Reference-Based Video Colorization/ColorMNet Video Colorization (New)
ComfyUI Node

ColorMNet Video Colorization (New)

A ComfyUI node in ColorMNet/Video with 10 inputs and 2 outputs.

By jonstreeter·Created 9 months ago·Updated 9 months ago· 24
ColorMNet Video Colorization (New)
  • video_frames
  • reference_image
  • colorized_frames
  • performance_report
target_width768
target_height432
memory_modebalanced
feature_encoderresnet50
use_fp16true
use_torch_compilefalse
text_guidance
text_guidance_weight0.30
CategoryColorMNet/Video

Inputs (10)

NameTypeDefaultDescription
video_framesIMAGEBatch of video frames to be colorized [N, H, W, 3]
reference_imageIMAGEColor reference image that provides the color palette
target_widthINT76864–2048Output width - must be multiple of 32 (will be adjusted automatically)
target_heightINT43264–2048Output height - must be multiple of 32 (will be adjusted automatically)
memory_modeCOMBObalancedMemory management strategy: 'low_memory' for large videos, 'high_quality' for best results, 'balanced' for general use
feature_encoderCOMBOresnet50Feature extraction model: resnet50 (ColorMNet default), vgg19 (fast), dinov2_vitb (recommended, 40-60% better), dinov2_vitl (best quality), clip_vitb (text-guided)
use_fp16BOOLEANtrueUse half-precision (FP16) for faster processing with lower VRAM usage (minimal quality impact)
use_torch_compileBOOLEANfalseEnable torch.compile optimization for 15-25% speedup (increases first-run compilation time)
text_guidanceoptSTRINGText prompt to guide colorization (only for clip_vitb encoder). Examples: 'warm sunset colors', 'vibrant anime style', 'cold winter landscape'
text_guidance_weightoptFLOAT0.300–1How much the text guidance influences colorization (0=no effect, 1=maximum effect)

Outputs (2)

NameTypeDescription
colorized_framesIMAGE
performance_reportSTRING