vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : suggest reduce ctx size when kv init fails

Browse source
This commit is contained in:
slaren 2024-11-02 00:55:19 +01:00
parent bf60f27cda
commit 20e12112fd
2 changed files with 2 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

1
src/llama.cpp
View file

@ -19520,6 +19520,7 @@ struct llama_context * llama_new_context_with_model(
if (!llama_kv_cache_init(ctx->kv_self, ctx, type_k, type_v, kv_size, cparams.offload_kqv)) {
LLAMA_LOG_ERROR("%s: llama_kv_cache_init() failed for self-attention cache\n", __func__);
LLAMA_LOG_ERROR("%s: suggestion: try using a smaller context size (-c command line option or llama_context_params.n_ctx)\n", __func__);
llama_free(ctx);
return nullptr;
}