thepatch/YuE2-3B-GGUF
Adapter vs m-a-p/SheetSage2. m-a-p/SheetSage2 is a LoRA adapter, not full weights: rank 64, on key_proj, out_proj, query_proj, value_proj in 24 layers (96 adapted matrices). It modifies thepatch/YuE2-3B-GGUF at load time.
Adapter vs m-a-p/SheetSage2. m-a-p/SheetSage2 is a LoRA adapter, not full weights: rank 64, on key_proj, out_proj, query_proj, value_proj in 24 layers (96 adapted matrices). It modifies thepatch/YuE2-3B-GGUF at load time.