tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-17 05:25:09 -04:00
Code Issues Packages Projects Releases Wiki Activity

sampling : avoid expensive softmax during greedy sampling (#9605)

Browse Source 
* sampling : avoid expensive softmax during greedy sampling

ggml-ci

* speculative : fix default RNG seed + set sparams.n_probs

* Update tests/test-sampling.cpp

Co-authored-by: slaren <slarengh@gmail.com>

* sampling : add clarifying comment [no ci]

---------

Co-authored-by: slaren <slarengh@gmail.com>
This commit is contained in:
Georgi Gerganov
2024-09-24 09:03:17 +03:00
committed by GitHub
parent c087b6f11d
commit b0f27361f3
5 changed files with 59 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
include/llama.h
View File

@@ -1066,6 +1066,7 @@ extern "C" {
LLAMA_API struct llama_sampler * llama_sampler_init_dist (uint32_t seed);
/// @details Sorts candidate tokens by their logits in descending order and calculate probabilities based on logits.
/// NOTE: Avoid using on the full vocabulary as the sorting can become slow. For example, apply top-k or top-p sampling first.
LLAMA_API struct llama_sampler * llama_sampler_init_softmax (void);
/// @details Top-K sampling described in academic paper "The Curious Case of Neural Text Degeneration" https://arxiv.org/abs/1904.09751