Hello.
I am a developer who is struggling to study local llm.
I have also installed ollama and open webui on Ubuntu, and I am learning various things.
Recently, I've been having fun finding various models and datasets. At first, I didn't know much about it, so I searched on Face Hugger, and there were times when I couldn't sleep because of the harsh content.
When I first started learning, I remembered that it was easy to find learning models on Face Hugger, but later I didn't know and ended up looking everywhere, even going to Face Hugger...
I still don't know why the name is HuggingFace, but thanks to Face Hugger, I learned it perfectly!
Have you had any funny experiences while searching for related information like me?
Next time, I'll try writing a challenge for open webui with 10 Tesla V100s.
Roughly summarized, for lightweight models (12B or less), one 5080 is faster than five V100s with 32GB of VRAM. Also, older models were not necessarily better performing just because they had more VRAM.
