Muse Glimmer 30B, Meta’s open agent model, and the second tiny model that guesses sixteen words ahead so it runs fast on one consumer graphics card

Ask

Ask about this presentation

Answers are generated from this presentation.

Chapters

1. 1Muse Glimmer 30B0:00
2. 2Meta squeezed a 30B model down to fit a 24 GB card0:09
3. 3It wins at using tools and loses at driving a desktop0:34
4. 4The second model fills sixteen blanks in one pass1:05
5. 5The big model checks all sixteen guesses in one pass1:27
6. 6The guesser reads five snapshots from inside the big model1:56
7. 73.1 times on an RTX 50902:33
8. 8Every word still has to pass the big model3:09