Neural Architecture Search On Nas Bench 201 1

Métriques

Accuracy (Test)

Accuracy (Val)

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	Accuracy (Test)	Accuracy (Val)	Paper Title	Repository
NAR	94.33	91.44	Generalized Global Ranking-Aware Neural Architecture Ranker for Efficient Image Classifier Search
β-DARTS	94.36	91.55	$β$-DARTS: Beta-Decay Regularization for Differentiable Architecture Search
DARTS-V2	54.30	39.77	-	-
GDAS	93.61	89.89	Searching for A Robust Neural Architecture in Four GPU Hours
arch2vec	94.18	91.41	Does Unsupervised Architecture Representation Learning Help Neural Architecture Search?
EPE-NAS (N=10)	92.63	89.90	EPE-NAS: Efficient Performance Estimation Without Training for Neural Architecture Search
SNAS	92.77	90.10	SNAS: Stochastic Neural Architecture Search
Shapley-NAS	94.37	91.61	Shapley-NAS: Discovering Operation Contribution for Neural Architecture Search
DARTS-	93.80	91.03	DARTS-: Robustly Stepping out of Performance Collapse Without Indicators
SETN	86.19	82.25	-	-
DARTS-V1	54.30	39.77	-	-
NAS-LID+RSPS	92.9	89.74	NAS-LID: Efficient Neural Architecture Search with Local Intrinsic Dimension
DU-DARTS	93.86	91.21	DU-DARTS: Decreasing the Uncertainty of Differentiable Architecture Search
AG-Net	94.37	91.61	Learning Where To Look -- Generative NAS is Surprisingly Efficient
BaLeNAS-TF	94.33	91.52	BaLeNAS: Differentiable Architecture Search via the Bayesian Learning Rule	-
Λ-DARTS	94.36	91.55	$Λ$-DARTS: Mitigating Performance Collapse by Harmonizing Operation Selection among Cells
GAEA DARTS (ERM)	94.1	-	Geometry-Aware Gradient Algorithms for Neural Architecture Search
RSPS	87.66	84.16	Random Search and Reproducibility for Neural Architecture Search
IS-DARTS	94.36	91.55	IS-DARTS: Stabilizing DARTS through Precise Measurement on Candidate Importance
CATCH-meta	-	91.33	CATCH: Context-based Meta Reinforcement Learning for Transferrable Architecture Search	-

0 of 37 row(s) selected.