You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
llama : wrap llama_new_context_with_model in try/catch
This fixes a crash where ggml_vk_allocate fails in llama_kv_cache_init,
but the exception is never caught.
Signed-off-by: Jared Van Bortel <[email protected]>
0 commit comments