Third-order self-attention

An experimental third-order multihead self-attention model (requires N^3 memory 😰). Details to come.

Name		Name	Last commit message	Last commit date
Latest commit History 9 Commits
test		test
util		util
.gitignore		.gitignore
README.md		README.md
gptmodel_with_third_order_attention.py		gptmodel_with_third_order_attention.py
models_mae.py		models_mae.py
third_order_attention.py		third_order_attention.py
train_mae.py		train_mae.py
train_mae.sh		train_mae.sh
vision_transformer_mae.py		vision_transformer_mae.py

Provide feedback