How much does convert-h5-to-ggml.py require on large models?

[github-bryld]

There is a newly trained icelandic model, I'd love to test out:
https://huggingface.co/language-and-voice-lab/whisper-large-icelandic-30k-steps-1000h/

When I try to use convert-h5-to-ggml.py, the process gets killed immediately.
This is in contrast to the small model at https://huggingface.co/Valdimarb13/whisper-small-icelandic/ which I can convert fine, but which unfortunately isn't very good.

I run Ubuntu as a virtual machine with 2 cores and 2 gigs of RAM.
Increasing to 4 gigs (recommended max for my system) doesn't change things.

Is my macbook simply too puny or is ther another explanation?
I couldn't git-clone the large model (the main model-bin took forever), so it's just a simple folder, if that is of importance.

[github-bryld]

Update: I re-tried this using Windows Subsystem For Linux on my main Windows machine and had the same result. I suppose the huggingface model's size of 6 GB is simply too much for my puny systems to process, but that "Killed" msg is not very enlightening.
Are any of you able to run this conversion and if so on what configurations? This may be my excuse for an upgrade.