Model Architectures For Training
#5
by
Taylor658
- opened
How effective are different model architectures, such as Transformers with co-attention mechanisms, at training with multimodal datasets like The Cauldron?
That's an interesting experimental question. That's unfortunately out of scope of the purpose of this dataset, but we hope that it will serve as a resource to answer that question!
VictorSanh
changed discussion status to
closed