]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: a5cacb2)
finetune : fix ggml_allocr lifetimes (tmp workaround) (#5033)
authorUzo Nweke <redacted>
Fri, 19 Jan 2024 18:20:50 +0000 (13:20 -0500)
committerGitHub <redacted>
Fri, 19 Jan 2024 18:20:50 +0000 (20:20 +0200)
commit381ee195721d8e747ee31a60c0751822b3072f02
treecdf6f6450351fbb0e7db12d6344c2f1d75acb50etree
parenta5cacb22b2114fd9adf61c00cbb237384d86bcedcommit | diff
finetune : fix ggml_allocr lifetimes (tmp workaround) (#5033)

* Fix issue with alloc causing max_compute_size to be calculated

* remove ggml_allocr_free as suggested in issue #4791
examples/train-text-from-scratch/train-text-from-scratch.cpp diff | blob | history
Packaging of ggml-org/llama.cpp