From 9b66378cacfb0a4e0401e7682ec66f6ab89348ed Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Tue, 10 Oct 2023 10:49:39 +0300 Subject: [PATCH] llama : print clam_kqv and max_alibi_bias hparams --- llama.cpp | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/llama.cpp b/llama.cpp index eaa008dbf..3b63b6401 100644 --- a/llama.cpp +++ b/llama.cpp @@ -2229,6 +2229,8 @@ static void llm_load_print_meta(llama_model_loader & ml, llama_model & model) { LLAMA_LOG_INFO("%s: n_gqa = %u\n", __func__, hparams.n_gqa()); LLAMA_LOG_INFO("%s: f_norm_eps = %.1e\n", __func__, hparams.f_norm_eps); LLAMA_LOG_INFO("%s: f_norm_rms_eps = %.1e\n", __func__, hparams.f_norm_rms_eps); + LLAMA_LOG_INFO("%s: f_clamp_kqv = %.1e\n", __func__, hparams.f_clamp_kqv); + LLAMA_LOG_INFO("%s: f_max_alibi_bias = %.1e\n", __func__, hparams.f_max_alibi_bias); LLAMA_LOG_INFO("%s: n_ff = %u\n", __func__, hparams.n_ff); LLAMA_LOG_INFO("%s: freq_base_train = %.1f\n", __func__, hparams.rope_freq_base_train); LLAMA_LOG_INFO("%s: freq_scale_train = %g\n", __func__, hparams.rope_freq_scale_train); @@ -5013,8 +5015,8 @@ static struct ggml_cgraph * llm_build_mpt( const int64_t n_embd_head = hparams.n_embd_head(); const int64_t n_embd_gqa = hparams.n_embd_gqa(); - const float norm_eps = hparams.f_norm_eps; - const float clamp_kqv = hparams.f_clamp_kqv; + const float norm_eps = hparams.f_norm_eps; + const float clamp_kqv = hparams.f_clamp_kqv; const float max_alibi_bias = hparams.f_max_alibi_bias; const int n_gpu_layers = model.n_gpu_layers;