server: rename --clear-idle to --cache-idle-slots (#21741)
This commit is contained in:
committed by
GitHub
parent
e365e658f0
commit
9d49acb2a7
+4
-4
@@ -1316,13 +1316,13 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
|
||||
}
|
||||
).set_env("LLAMA_ARG_KV_UNIFIED").set_examples({LLAMA_EXAMPLE_SERVER, LLAMA_EXAMPLE_PERPLEXITY, LLAMA_EXAMPLE_BATCHED, LLAMA_EXAMPLE_BENCH, LLAMA_EXAMPLE_PARALLEL}));
|
||||
add_opt(common_arg(
|
||||
{"--clear-idle"},
|
||||
{"--no-clear-idle"},
|
||||
{"--cache-idle-slots"},
|
||||
{"--no-cache-idle-slots"},
|
||||
"save and clear idle slots on new task (default: enabled, requires unified KV and cache-ram)",
|
||||
[](common_params & params, bool value) {
|
||||
params.clear_idle = value;
|
||||
params.cache_idle_slots = value;
|
||||
}
|
||||
).set_env("LLAMA_ARG_CLEAR_IDLE").set_examples({LLAMA_EXAMPLE_SERVER}));
|
||||
).set_env("LLAMA_ARG_CACHE_IDLE_SLOTS").set_examples({LLAMA_EXAMPLE_SERVER}));
|
||||
add_opt(common_arg(
|
||||
{"--context-shift"},
|
||||
{"--no-context-shift"},
|
||||
|
||||
Reference in New Issue
Block a user