Remove scary warning about incompatible model

This commit is contained in:
Iwan Kawrakow
2025-08-22 18:15:05 +03:00
parent dfa6e2b5fa
commit 9fa5956602

View File

@@ -4072,22 +4072,21 @@ static void llm_load_hparams(
case LLM_ARCH_DEEPSEEK2:
{
if (hparams.n_head_kv() == 1) {
printf("==========================================================================\n");
printf("Detected incompatible DeepSeek model.\n");
printf("Will try to fix, but there are no guarantees\n\n");
printf("*** Your prompt processing speed will be crippled ***\n\n");
printf("Consider making your own ik_llama.cpp compatible model or\n");
printf("ask the model provider to make one for you,\n");
int n_nead_kv = hparams.n_gqa();
if (n_nead_kv%16 != 0 || hparams.n_embd_head_k != 576 || hparams.n_embd_head_v != 512 ||
hparams.n_rot != 64) {
printf("==========================================================================\n");
printf("Detected incompatible DeepSeek model without a known way to fixc it.\n");
printf("Consider making your own ik_llama.cpp compatible model or\n");
printf("ask the model provider to make one for you,\n\n");
printf("Sorry, uknown model => cannot fix it => bailing out\n");
printf("==========================================================================\n");
GGML_ABORT("Fatal error");
}
printf("================= Adjusted mainline llama.cpp MLA tensors to ik_llama.cpp\n");
for (auto& item : hparams.n_head_kv_arr) item = n_nead_kv;
hparams.n_embd_head_k = 192;
hparams.n_embd_head_v = 128;
printf("==========================================================================\n");
//GGML_ABORT("Fatal error");
}
bool is_lite = (hparams.n_layer == 27);