Because the uploader obviously had no idea how to properly choose a DIM and rank when training. Just run this LoRA through Kohya to trim the fat. Trimmed down to 96MB, mine still works just fine.

@Bit_Shifter can i hasthat