tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-29 20:45:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : add check for KV cache shifts (#10401)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2024-11-19 13:29:26 +02:00
committed by GitHub
parent a88ad007de
commit 8e752a777b
3 changed files with 14 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
common/common.cpp
View File

@ -875,6 +875,12 @@ struct common_init_result common_init_from_params(common_params & params) {
return iparams;
}
if (params.ctx_shift && !llama_kv_cache_can_shift(lctx)) {
LOG_ERR("%s: KV cache shifting is not supported for this model (--no-context-shift to disable)'\n", __func__);
llama_free_model(model);
return iparams;
}
if (!params.control_vectors.empty()) {
if (params.control_vector_layer_start <= 0) params.control_vector_layer_start = 1;
if (params.control_vector_layer_end <= 0) params.control_vector_layer_end = llama_n_layer(model);