]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 098dbaa)
perplexity : support using multiple sequences to allow larger batch sizes (#5946)
authorslaren <redacted>
Sat, 9 Mar 2024 18:55:54 +0000 (19:55 +0100)
committerGitHub <redacted>
Sat, 9 Mar 2024 18:55:54 +0000 (19:55 +0100)
commitd894f352bf433157232dc8dc54eacd50014e898e
tree0ed89f222eb3f0e08a397c953d27e7e6846156cetree
parent098dbaab449f5309a54871ba7e5acef72ae696decommit | diff
perplexity : support using multiple sequences to allow larger batch sizes (#5946)

* perplexity : support using multiple sequences to allow larger batch sizes

ggml-ci

* set cparams.n_parallel to the number of sequences

* print tested n_ctx, add assert
examples/perplexity/perplexity.cpp diff | blob | history
llama.cpp diff | blob | history
Packaging of ggml-org/llama.cpp