MCPcopy Create free account
hub / github.com/appdevforall/CodeOnTheGo / n_embd_k_gqa

Method n_embd_k_gqa

subprojects/llama.cpp/src/llama-hparams.cpp:79–83  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

77}
78
79uint32_t llama_hparams::n_embd_k_gqa(uint32_t il) const {
80 const uint32_t n_head_kv = this->n_head_kv(il);
81
82 return n_embd_head_k * n_head_kv;
83}
84
85uint32_t llama_hparams::n_embd_v_gqa(uint32_t il) const {
86 const uint32_t n_head_kv = this->n_head_kv(il);

Callers 9

load_tensorsMethod · 0.80
print_infoMethod · 0.80
llm_graph_contextMethod · 0.80
llama_kv_cacheMethod · 0.80
get_kMethod · 0.80
build_graph_shiftMethod · 0.80
state_write_dataMethod · 0.80
state_read_dataMethod · 0.80
build_attn_blockMethod · 0.80

Calls 1

n_head_kvMethod · 0.95

Tested by

no test coverage detected