/hard-attention

Leaning hard attention model by policy gradient with rewards based on active inference.

Primary LanguageLuaMIT LicenseMIT

No issues in this repository yet.