mirror of
https://github.com/likelovewant/ollama-for-amd.git
synced 2025-12-21 22:33:56 +00:00
When converting a ggml model if there is a failure to read tensor data a nil error value was being returned. It should be assigned to the actual error from reading.