Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions common/arg.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -3163,7 +3163,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
params.speculative.cpuparams.n_threads = std::thread::hardware_concurrency();
}
}
).set_examples({LLAMA_EXAMPLE_SPECULATIVE}));
).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}));
add_opt(common_arg(
{"-tbd", "--threads-batch-draft"}, "N",
"number of threads to use during batch and prompt processing (default: same as --threads-draft)",
Expand All @@ -3173,7 +3173,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
params.speculative.cpuparams_batch.n_threads = std::thread::hardware_concurrency();
}
}
).set_examples({LLAMA_EXAMPLE_SPECULATIVE}));
).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}));
add_opt(common_arg(
{"-Cd", "--cpu-mask-draft"}, "M",
"Draft model CPU affinity mask. Complements cpu-range-draft (default: same as --cpu-mask)",
Expand Down
2 changes: 2 additions & 0 deletions tools/server/server.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -2015,6 +2015,8 @@ struct server_context {
params_dft.cache_type_k = params_base.speculative.cache_type_k;
params_dft.cache_type_v = params_base.speculative.cache_type_v;

params_dft.cpuparams.n_threads = params_base.speculative.cpuparams.n_threads;
params_dft.cpuparams_batch.n_threads = params_base.speculative.cpuparams_batch.n_threads;
params_dft.tensor_buft_overrides = params_base.speculative.tensor_buft_overrides;

llama_init_dft = common_init_from_params(params_dft);
Expand Down
Loading