AWhat changed
Jointly optimized training speed and parameter efficiency through architecture search/scaling and adjusted regularization as training resolution increased.
Scientific scope
Jointly optimized training speed and parameter efficiency through architecture search/scaling and adjusted regularization as training resolution increased. Training-aware architecture search selected network designs within a human-defined space; researchers also supplied scaling and progressive-training choices. Inherits EfficientNet, fused mobile convolutions and progressive resizing. Delta is the training-aware integrated design and adaptive regularization, not a model-size version. Relative training gains depend on resolution, hardware, data and baseline.
Evidence
Training-aware architecture search selected network designs within a human-defined space; researchers also supplied scaling and progressive-training choices. WRS has not independently rerun these experiments. Training-aware architecture search selected network designs within a human-defined space; researchers also supplied scaling and progressive-training choices. Inherits EfficientNet, fused mobile convolutions and progressive resizing. Delta is the training-aware integrated design and adaptive regularization, not a model-size version. Relative training gains depend on resolution, hardware, data and baseline.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction