coriander VS sparsegpt

It’s not a full implementation of CUDA and requires some contortions to use but https://github.com/hughperkins/coriander is as good as anything else I’ve tried. It has been a few years though.

It’s possible to run cuda on anything . There have been attempts to do this. https://github.com/hughperkins/coriander Unfortunately it seems development stalled.

SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot (https://arxiv.org/abs/2301.00774)

There is this : https://github.com/IST-DASLab/sparsegpt

Check the paper here, it's intersting: https://arxiv.org/abs/2301.00774

There's no chance that we've peeked from a bang for buck sense - we still haven't adequately investigated sparse networks.

Relevantish: https://arxiv.org/abs/2301.00774

The fact that we can reach those levels of sparseness with pruning also indicates that we're not doing a very good job of generating the initial network conditions.

Being able to come up with trainable initial settings for sparse networks across different topologies is hard, but given that we've had a degree of success with pre-trained networks, pre-training and pre-pruning might also allow for sparse networks with minimally compromised learning capabilities.

If it's possible to pre-train composable network modules, it might also be feasible to define trainable sparse networks with significantly relaxed topological constraints.

Training uses gradient descent, so you want to have good precision during that process. But once you have the overall structure of the network, https://arxiv.org/abs/2210.17323 (GPTQ) showed that you can cut down the precision quite a bit without losing a lot of accuracy. It seems you can cut down further for larger models. For the 13B Llama-based ones, going below 5 bit per parameter is noticeably worse, but for 30B models you can do 4 bits.

The same group did another paper https://arxiv.org/abs/2301.00774 which shows that in addition to reducing the precision of each parameter, you can also prune out a bunch of parameters entirely. It's harder to apply this optimization because models are usually loaded into RAM densely, but I hope someone figures out how to do it for popular models.