CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-27 08:21:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : remove KV cache defragmentation logic (#15473)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-08-22 12:22:13 +03:00
committed by GitHub
parent ad5c975c2d
commit 9ebebef62f
16 changed files with 32 additions and 440 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
common/arg.cpp
View File

@@ -2254,9 +2254,11 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
).set_examples({LLAMA_EXAMPLE_PERPLEXITY}));
add_opt(common_arg(
{"-dt", "--defrag-thold"}, "N",
string_format("KV cache defragmentation threshold (default: %.1f, < 0 - disabled)", (double)params.defrag_thold),
string_format("KV cache defragmentation threshold (DEPRECATED)"),
[](common_params & params, const std::string & value) {
params.defrag_thold = std::stof(value);
GGML_UNUSED(params);
GGML_UNUSED(value);
LOG_WRN("DEPRECATED: --defrag-thold is deprecated and no longer necessary to specify\n");
}
).set_env("LLAMA_ARG_DEFRAG_THOLD"));
add_opt(common_arg(