Comfyui workflow
Could you provide a ComfyUI workflow example? For instance, loading model weights with LoRA while removing VAE loading — is this the correct way to use it?
https://github.com/alcaitiff/ComfyUI-AsymFLUX install the custom node and the workflow is on the folder workflows
ive implemented gguf support and fp4/fp8 encoding for low vram if needed. currently testing before i update the github fork i made. currently only holds bf16 model and fp4 and fp8 encoders but runs on 8gb vram and 16gb ram for me so ill update when gguf implmentation is tested fully.
https://github.com/RealRebelAI/REBELS_LOW_VRAM_ComfyUI-Klein9B-AsymFlow/tree/main
Here is my implementation - no extra dependencies required, works with Comfy's KSampler
https://github.com/Nynxz/ComfyUI-AsymFLUX2
https://github.com/alcaitiff/ComfyUI-AsymFLUX install the custom node and the workflow is on the folder workflows
I’m using a GGUF‑quantized model, so please merge it into your main branch once @realrebelai adds GGUF support. Thank you.
https://github.com/alcaitiff/ComfyUI-AsymFLUX install the custom node and the workflow is on the folder workflows
I’m using a GGUF‑quantized model, so please merge it into your main branch once @realrebelai adds GGUF support. Thank you.
Mine works with GGUF models if you want to try it out in the meantime!
Here is my implementation - no extra dependencies required, works with Comfy's KSampler
I actually had to delete mine an fork yours because theres something very off about how claude coded mine. Usually not a problem but it went a weird route with the other nodes i forked. I added a pixel smoothing node if you want to pull from it and throw it in yours!