LLM-Inference/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-08-12 22:29:39 +04:00
Code Issues Packages Projects Releases Wiki Activity

fix: initialize draft model parameters with base values (#2178)

Browse Source
This commit is contained in:
Samuel Oliveira Alves
2026-07-24 18:11:17 +03:00
committed by GitHub
parent 31018dc511
commit f359df4bc9
1 changed files with 1 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files
common/speculative.cpp
+1 -19
View File
@@ -1884,30 +1884,13 @@ bool common_speculative_load_draft_model(
return true;
}
gpt_params params_dft;
gpt_params params_dft = params_base;
params_dft.devices = params.devices;
params_dft.model = params.model;
params_dft.main_gpu = params_base.main_gpu;
params_dft.n_gpu_layers = params.n_gpu_layers;
params_dft.rpc_servers = params_base.rpc_servers;
params_dft.cache_type_k = params.cache_type_k.empty() ? params_base.cache_type_k : params.cache_type_k;
params_dft.cache_type_v = params.cache_type_v.empty() ? params_base.cache_type_v : params.cache_type_v;
params_dft.flash_attn = params_base.flash_attn;
params_dft.k_cache_hadamard = params_base.k_cache_hadamard;
params_dft.v_cache_hadamard = params_base.v_cache_hadamard;
if (params.has_stage_type(COMMON_SPECULATIVE_TYPE_DFLASH)) {
params_dft.split_mode = params_base.split_mode;
for (size_t i = 0; i < std::size(params_dft.tensor_split); ++i) {
params_dft.tensor_split[i] = params_base.tensor_split[i];
}
params_dft.attn_max_batch = params_base.attn_max_batch;
params_dft.graph_reuse = params_base.graph_reuse;
params_dft.split_mode_graph_scheduling = params_base.split_mode_graph_scheduling;
params_dft.scheduler_async = params_base.scheduler_async;
params_dft.max_extra_alloc_MiB = params_base.max_extra_alloc_MiB;
params_dft.reduce_type = params_base.reduce_type;
}
if (!params.params.empty()) {
auto [argc, argv] = parse_command_line("llama-server " + params.params);
@@ -1929,7 +1912,6 @@ bool common_speculative_load_draft_model(
}
params_dft.n_ctx = params_dft.n_ctx == 0 ? params_base.n_ctx / params_base.n_parallel : params_dft.n_ctx;
params_dft.n_parallel = 1;
params_dft.n_batch = params_dft.n_ctx;
params.mparams_dft.path = params_dft.model;