Skip to content

Conversation

ggerganov
Copy link
Member

The graph reuse logic (#14482) was broken because the graph inputs could hold dangling references to llama_hparams and llama_cparams. This was spotted after enabling LLAMA_SET_ROWS=1 in the CI: #14959. The error is here: https://github.com/ggml-org/llama.cpp/actions/runs/16615559549/job/47007325550?pr=14959#step:7:21495

@ggerganov ggerganov merged commit 1e15bfd into master Jul 30, 2025
55 checks passed
@ggerganov ggerganov deleted the gg/graph-fix-stack-use-after-return branch July 30, 2025 10:52
Nexesenex pushed a commit to Nexesenex/croco.cpp that referenced this pull request Aug 1, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant