Basics

• gguf naming convention: https://github.com/ggml-org/ggml/blob/master/docs/gguf.md

• safetensors doc: https://hf-mirror.com/docs/safetensors/index

• save_gguf

• save_safetensors

• savez

mean

• torch: keepdim

• mlx: keepdims

• torch: dim

• mlx: axis

cat

torch uses cat, while mlx uses concatenate.

torch uses dim in cat, mlx uses axis in concatenate

Sequential

torch users this_layer[0], while mx uses this_layer.layers[0]