llama : document logits_all deprecation

llama_context_params.logits_all is a parameter for controlling
llama_eval. This documents that logits_all should not be used with
llama_decode and llama_batch.
This commit is contained in:
crasm 2023-12-11 17:53:02 -05:00
parent 8a7b2fa528
commit a13ba22a3c

View file

@ -216,7 +216,7 @@ extern "C" {
// Keep the booleans together to avoid misalignment during copy-by-value. // Keep the booleans together to avoid misalignment during copy-by-value.
bool mul_mat_q; // if true, use experimental mul_mat_q kernels (DEPRECATED - always true) bool mul_mat_q; // if true, use experimental mul_mat_q kernels (DEPRECATED - always true)
bool logits_all; // the llama_eval() call computes all logits, not just the last one bool logits_all; // the llama_eval() call computes all logits, not just the last one (DEPRECATED - set llama_batch.logits instead)
bool embedding; // embedding mode only bool embedding; // embedding mode only
bool offload_kqv; // whether to offload the KQV ops (including the KV cache) to GPU bool offload_kqv; // whether to offload the KQV ops (including the KV cache) to GPU
}; };