LLM inference in C/C++
C++ ggml
model: load YaRN beta metadata for laguna models (#30331)
Adam Caldwellf441d2d9db68
diff --git a/src/models/laguna.cpp b/src/models/laguna.cppindex b59d8afc7a..cbc7b26d7c 100644--- a/src/models/laguna.cpp+++ b/src/models/laguna.cpp@@ -7,6 +7,9 @@#include "models.h"void llama_model_laguna::load_arch_hparams(llama_model_loader & ml) {+ ml.get_key(LLM_KV_ROPE_SCALING_YARN_BETA_FAST, hparams.yarn_beta_fast, false);+ ml.get_key(LLM_KV_ROPE_SCALING_YARN_BETA_SLOW, hparams.yarn_beta_slow, false);+ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);ml.get_key(LLM_KV_LEADING_DENSE_BLOCK_COUNT, hparams.n_layer_dense_lead);ml.get_key_or_arr(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp_arr, hparams.n_layer_all);