Run Taggui on cloud gpu

[BenDes21]

Hi there, would like to thanks again @jhc13 for this amazing tools, really nice job.

As you know, some of the best LLM models like Cogvlm V2 require a lot of vram ( no 4 bit version )

I would like to know, for people for don't have the necessary resource's for run theses models, if they're is a way a use taggui into a cloud service who can support high vram gpu, heard that unfortunately create a runpod template was really hard for taggui.

Thanks

[geroldmeisinger]

for all VLMs which run on llama.cpp you might have luck with Ollama. CogVLM2 is unfortunately not supported: ollama/ollama#1930

[jhc13]

TagGUI requires a graphical interface to run. You can rent a cloud server and install a desktop environment on it (example).

[BenDes21]

thanks