Why isnt LLM training itself open sourced? With all the compute in the world, something like Folding@home here would be killer
data bandwidth limits distributed training under current architectures. really interesting implications if we can make progress on that
It's either illegal or extremely expensive to source quality training material.
Well it is, it's in the name "OpenAI". /S
It is in some cases. NVIDIA's models are open source, in the truest sense that you can download the training set and training scripts and make your own.