llama: various bug fixes (#26051)

This commit is contained in:
Xuan-Son Nguyen
2026-07-24 18:56:42 +02:00
committed by GitHub
parent fa72aeccb2
commit 298219f985
7 changed files with 46 additions and 3 deletions
+1
View File
@@ -1355,6 +1355,7 @@ llama_model * llama_quant_model_from_metadata(const llama_quant_model_desc * des
model->hparams.n_embd_head_k_full = desc->n_embd_head_k;
model->hparams.n_embd_head_v_full = desc->n_embd_head_v;
model->hparams.n_layer_all = desc->n_layer;
GGML_ASSERT(desc->n_layer > 0 && desc->n_layer <= LLAMA_MAX_LAYERS);
model->hparams.n_expert = desc->n_expert;
for (uint32_t i = 0; i < desc->n_layer; i++) {