create_module built each up/down module with the default constructor,
which kaiming-initializes the parameter, then copied the stored weight
over the whole thing. The init is thrown away every time and costs about
four times the copy: 22.1ms per module against 2.5ms, or 5.8s against
0.7s over a 264-module lora, on every load.
skip_init constructs on meta and materializes uninitialized, so the copy
still fully defines the parameter. Dtype, device and values are
unchanged, including the fp32 upcast of bf16 files that the copy performs.
Route nn.Embedding targets (and the SDNQEmbedding / ScaledWordEmbedding subclasses) through the linear LoRA path: the weight delta is up@down over the [vocab, dim] table, same shape and merge as a Linear.
Apply a companion bias delta (diff_b) as ex_bias on the same module rather than dropping it; collect diff_b into the LoRA group so it rides the existing module instead of a separate Full module that would collide on the network key.
CalamitousFelicitousness
Disty0
Vladimir Mandic
AI-Casanova