fix: step35 MTP does not allocate KV cache for all layers - #24125

fix: step35 MTP does not allocate KV cache for all layers#24125

pwilkin merged 1 commit into

ggml-org:masterggml-org/llama.cpp:masterfrom

forforever73:step35-mtp-kv-filterforforever73/llama.cpp:step35-mtp-kv-filterCopy head branch name to clipboard