Jesse Gross cec8a9dee0 ollamarunner: Re-enable worst case graph preallocation. ... Worst case graph preallocation was disabled by a27462b "ollamarunner: Temporarily disable worst case graph preallocation" since it caused crashes with large batches when not using the GPU. This backports upstream llama.cpp commit f057808 "ggml: Don't assert fail when tensor data changes (#13222)", which fixes the underlying bug and allows reverting the previous workaround.		2025-12-29 06:37:50 -06:00
..
backend	ollamarunner: Re-enable worst case graph preallocation.	2025-12-29 06:37:50 -06:00
nn	attention: Remove unnecessary contiguous operations	2025-03-01 20:53:23 -08:00
backend.go	llama4	2025-12-29 06:37:44 -06:00