A transformer next-frame predictor that feeds each attention head the full embedding and trains on embedding-space loss reports gains on three video datasets, but the comparison is confounded by using the training objective as the test metric.
Anticipative video transformer
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Overcoming Semantic Dilution in Transformer-Based Next Frame Prediction
A transformer next-frame predictor that feeds each attention head the full embedding and trains on embedding-space loss reports gains on three video datasets, but the comparison is confounded by using the training objective as the test metric.