Skip to content

server : fix default draft model parameters #10586

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Dec 3, 2024
Merged

server : fix default draft model parameters #10586

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
150d6e9
server : force F16 KV cache for the draft model
ggerganov Nov 29, 2024
f325205
server : fix draft params
ggerganov Dec 2, 2024
11b4d58
server : various params fixes
ggerganov Dec 3, 2024
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
17 changes: 14 additions & 3 deletions examples/server/server.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -694,8 +694,9 @@ struct server_context {

params_dft.devices = params_base.speculative.devices;
params_dft.model = params_base.speculative.model;
params_dft.n_ctx = params_base.speculative.n_ctx;
params_dft.n_ctx = params_base.speculative.n_ctx == 0 ? params_base.n_ctx / params_base.n_parallel : params_base.speculative.n_ctx;
params_dft.n_gpu_layers = params_base.speculative.n_gpu_layers;
params_dft.n_parallel = 1;

common_init_result llama_init_dft = common_init_from_params(params_dft);

Expand All @@ -715,8 +716,14 @@ struct server_context {
return false;
}

cparams_dft = common_context_params_to_llama(params_base);
cparams_dft.n_batch = llama_n_ctx(llama_init_dft.context);
const int n_ctx_dft = llama_n_ctx(llama_init_dft.context);

cparams_dft = common_context_params_to_llama(params_dft);
cparams_dft.n_batch = n_ctx_dft;

// force F16 KV cache for the draft model for extra performance
cparams_dft.type_k = GGML_TYPE_F16;
cparams_dft.type_v = GGML_TYPE_F16;

// the context is not needed - we will create one for each slot
llama_free(llama_init_dft.context);
Expand Down Expand Up @@ -2308,6 +2315,10 @@ struct server_context {
continue;
}

if (slot.state != SLOT_STATE_GENERATING) {
continue;
}

llama_token id = slot.sampled;

struct common_speculative_params params_spec;
Expand Down
Loading