Comfyui Workflow (Image Generation)
required Custom Nodes:
ComfyUI_Eclipse: Version 2.3.13, ComfyReg (Nodes 2.0 compat, V3 API)
ComfyUI_SmartLML: Version 2.3.14, ComfyReg (Nodes 2.0 compat, V3 API)
added fast bypasser etc. from rgthree to eclipse (v3, nodes 2.0 compat)
if you need help installing nunchaku or how to use the nodes check the readme files in the repo readme folder at github.
llama-cpp-python is needed for the loading_method: GGUF (llama-cpp-python) in the Smart LML Node
nunchaku wheels:
nunchaku model files:
Wildcards
Description
reuploaded with the following change:
added smart loader plus to the detailers this allows you to load a template, set allow_overwrite in the sampler settings node to yes and use the saved values. its important to switch back the template action from load to none to keep the changes for e.g. step values when reloading or loading the workflow, there is also no need to use the configure_latent option in the detailers. if the model loader is disabled and allow_overwrite is still true the values from the initial render are used for the detailing (not recommended to do the full steps with a low denoise, except with lightspeed maybe^^).The name of the node was changed from “rvtools” to “eclipse” because problems arose when using ‘rvtools_v2’ at the same time, and the old name gave the impression that “rvtools_v2” was newer. I apologize for any inconvenience this may have caused.
update eclipse (1.0.30)
fixed some connections in the inital render
added qwen image diffsynth to the controlnet group
changed the detailer by adding a subgraph for the detection / sam2 to clean it up with only 1 stop button now