PagedLion8bit crashes NVIDIA GPU when saving state?

[ChiNoel-osu]

It's probably just me because there's literally no information about my issue. But I'll start a discussion anyway hopefully we can leave something on the internet.
I have a 4060 laptop (8Gi) in MSHybrid mode (dGPU do rendering and iGPU outputs), Studio driver 572.60. When training it always exceeds the physical VRAM and will use shared VRAM (offloads to DRAM). Thus making the training very slow and also low GPU usage.
The speed is not the problem though, the issue is when it finishes an epoch and saves state, the save will never happen, and the script just stucks. Event Viewer shows nvlddmkm event ID 14 that reads:

The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event: 

\Device\UVMLiteProcess1
GPU recovery action changed from 0x0 (None) to 0x2 (Node Reboot Required)

The message resource is present but the message was not found in the message table

It will do that every single time. However the normal Lion8bit optimizer works just fine.

I assume using the shared VRAM is what causes the issue? I have no idea how PagedLion8bit works so if anyone could shed some light on this I'm very appreciated.