[edude03]

Maybe the only downside to how fast LLMs are moving is papers come out faster than anyone (not at Google) can train and test the improvements.

I got into deep learning around when ReLU and dropout was hot and on my consumer 1080 I was able to change one or two lines of code and test the improvements in a few hours, whereas now, I guess I'll need to wait a few weeks for mistral et al to try it out

[htrp]

Welcome to the GPU poor!

I'm focusing in quantization approaches and testing on my obsolete last gen GPUs.

[edude03]

The funny thing is, I have 8 3090s which last epoch would have put in like - top 1% of compute. Now, still a lot of compute but pales in comparison to the 100x H100 GPU clusters we're seeing today.