ColoringBook - KontextDev - 1 I2I - 1 Lora - V2 — v2.0

ColoringBook - KontextDev - 1 I2I - 1 Lora - V2 — v2.0 is for transforming an uploaded image into a drawing ready to be colorized; its input is an image and its output is that drawing.

Version v2.0 adds LORA support, a Prompt SELECTER, and a PRE-VAE-Encoder UPScaler for more details.

The workflow description lists one Guidance CFG, the Euler A sampler, and a shift setting from 0.2 to 1.

Input: an uploaded image. Output: a drawing ready to be colorized.

It uses Flux.1 Kontext as the base model and is published by OleaSTeR.

Named model files are ae.safetensors, clip_l.safetensors, and t5xxl_fp8_e4m3fn_scaled.safetensors.

Named workflow components include FluxKontextImageScale, ReferenceLatent, CheckpointLoaderSimple, Image Size to Number, PrimitiveStringMultiline, DualCLIPLoader, VAEEncode, CR Text Input Switch (4 way), SDXLEmptyLatentSizePicker+, and VAEDecode.

Other named components are Lora Loader, LoadImage, easy imageSave, CLIPSetLastLayer, VAELoader, ImageScaleToTotalPixels, FluxTextSampler, and CLIPTextEncode.

The listed background choice is No background.

The listed application settings are final image size equal to source image size, batch 1, 10 steps, guidance 2.5, denoise 1, and clipskip 2.

Before using it, verify that ae.safetensors and t5xxl_fp8_e4m3fn_scaled.safetensors are the intended files.

Vorschlag · nicht geprüft

The listing does not specify a model-use license; check the publisher's usage terms before use.

Vorschlag · nicht geprüft