Are there any expansion plans for comfyui？

[shiyk92]

ty

[pharmapsychotic]

oh good idea! I've been doing bunch of other stuff with comfy lately

[hben35096]

Expect it to appear

[TomLucidor]

Seconded as well even if WD14 and DeepDanbooru exists, need to be retrofitted with BLIP and BLIP2 as well without the need of initial prompts

• WaifuDiffusion? https://github.com/pythongosssss/ComfyUI-WD14-Tagger
• Deepdanbooru (classic) https://github.com/gamert/ComfyUI_tagger
• BLIP-based but weird https://github.com/WASasquatch/was-node-suite-comfyui
• GIT/COCA? https://github.com/theovercomer8/captionr
• Whatever this is (don't roast me bro) https://github.com/abhishekkrthakur/diffuzers

Side note: this exists as well, but probably needs longer (and more detailed) hard prompts https://github.com/YuxinWenRick/hard-prompts-made-easy

Cross-reference with https://github.com/comfyanonymous/ComfyUI/issues/990

[formulake]

Just to revive the subject. This approach to captioning is really solid. The ComfyUI implementation would be a runaway success.

[hben35096]

Just to revive the subject. This approach to captioning is really solid. The ComfyUI implementation would be a runaway success. There are already there https://github.com/shadowcz007/comfyui-mixlab-nodes

[formulake]

Just to revive the subject. This approach to captioning is really solid. The ComfyUI implementation would be a runaway success. There are already there https://github.com/shadowcz007/comfyui-mixlab-nodes

Thanks a MILLION!

[TomLucidor]

@formulake would this be supported in Flux and AuraFlow with how they toy with LLMs? Hope that it will work tho