aukejw

I'm also trusting `get_peak_memory` + some small buffer for now.

Still, it reports accurate peak memory usage for tensors living on GPU, but seems to miss some of the non-Metal overhead, however small (https://github.com/aukejw/mlx_transformers_benchmark/issues/...).

How do you determine peak memory usage? Just look at activity monitor?

I've yet to find a good overview of how much memory each model needs for different context lengths (other than back of the envelope #weights * bits). LM Studio warns you if a model will likely not fit, but it's not very exact.

There are plenty of smaller (quantized) models that fit well on your machine! On a M4 with 24GB it’s already possible to comfortably run 8B quantized models.

Im benchmarking runtime and memory usage for a few of them: https://aukejw.github.io/mlx_transformers_benchmark/