- We need to train Hyperbolic Adapters - We should try to do this with CLIP embeddings first - Likely requires a dataset of Hierarchical Data
We need to train Hyperbolic Adapters
We should try to do this with CLIP embeddings first
Likely requires a dataset of Hierarchical Data