ill have to retry my llamagpt because whenever i use it it goes slower than both screens. seems to be a ram issue with llama using sub 1gb of RAM on my system
In case you're wondering how much more powerful the new Umbrel Home is vs Pi 4, the difference is in orders of magnitude:
- 100% Bitcoin node sync: 6x faster (in just ~6.5 hours)
- File compression: 9.5x faster
- Plex video transcoding: 3.8x faster
- LlamaGPT: 5x faster
Here’s a demo of LlamaGPT interference and generation happening locally on both the devices:
https://video.nostr.build/0e3b1c3ab8abcd17fcc9589dc6ed39f7db8bf1817d77b60e69e84c8d18f136e1.mp4
Discussion
No replies yet.