Models
JiT-T2I-XXL
A strong late-fusion text-to-image baseline that Halcyon outperformed with far less compute.
0 stories.
Every story about JiT-T2I-XXL (0)
No story is about JiT-T2I-XXL yet: the ones below only mention it.
Also mentioned in (1)
Stories about something else that name JiT-T2I-XXL in passing, with the sentence that does.
- Image Models Trained Without Language Pretraining Matched Rivals in a Controlled Test“An even larger Halcyon model, with 1.6 billion parameters, beat another strong late-fusion baseline called JiT-T2I-XXL while using 275 times less pretraining compute.”