]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 65bdd52)
#1869 Fix null reference errors when training from scratch with CUDA (#1907)
authorRobyn <redacted>
Sat, 24 Jun 2023 18:10:29 +0000 (04:10 +1000)
committerGitHub <redacted>
Sat, 24 Jun 2023 18:10:29 +0000 (20:10 +0200)
commit5ec8dd5a3c6a9a109351d2257bb9d53869bd0a94
tree571f9f195508ff20b7d23aa698c2660d0d0f8ab5tree
parent65bdd52a867539691007f85c5508146d507f72c1commit | diff
#1869 Fix null reference errors when training from scratch with CUDA (#1907)

* #1869 Fix null reference errors when training from scratch with CUDA build

Calling ggml_compute_forward when node->src0 was null was causing train-text-from-scratch.exe to terminate unexpectedly.

* ggml : do not dereference src0 if NULL

---------

Co-authored-by: Georgi Gerganov <redacted>
ggml-cuda.cu diff | blob | history
ggml.c diff | blob | history
Packaging of ggml-org/llama.cpp