llama.cpp

quantize: Handle user-defined pruning of whole layers (blocks)

#13037

Merged