HyperAIHyperAI超神经

Command Palette

Search for a command to run...

算力平台

Neural Architecture Search On Imagenet

评估指标

Accuracy
FLOPs
Params
Top-1 Error Rate

评测结果

各个模型在此基准测试上的表现结果

Paper Title
MUXNet-xs66.7-1.8M33.3MUXConv: Information Multiplexing in Convolutional Neural Networks
FBNetV2-F168.3--31.7FBNetV2: Differentiable Neural Architecture Search for Spatial and Channel Dimensions
LayerNAS-60M--3.7M31LayerNAS: Neural Architecture Search in Polynomial Complexity
NASGEP70.49--29.51Optimizing Neural Architecture Search using Limited GPU Time in a Dynamic Search Space: A Gene Expression Programming Approach
MUXNet-s71.6-2.4M28.4MUXConv: Information Multiplexing in Convolutional Neural Networks
NN-MASS-A72.9200M2.3M27.1How does topology of neural architectures impact gradient propagation and model performance?
FBNetV2-F373.2--26.8FBNetV2: Differentiable Neural Architecture Search for Spatial and Channel Dimensions
DARTS73.3-4.926.7DARTS: Differentiable Architecture Search
NN-MASS-B73.3393M3.7M26.7How does topology of neural architectures impact gradient propagation and model performance?
GDAS74.0-5.326.0-
MemNAS74.1--25.9MemNAS: Memory-Efficient Neural Architecture Search With Grow-Trim Learning
EEEA-Net-C1 (b=7) (our)74.3--25.7EEEA-Net: An Early Exit Evolutionary Neural Architecture Search
FairNAS-C74.69-4.4M25.4FairNAS: Rethinking Evaluation Fairness of Weight Sharing Neural Architecture Search
RandWire-WS (small)-583M5.6M25.3Exploring Randomly Wired Neural Networks for Image Recognition
iDARTS (CIFAR-10)--5.1M25.2iDARTS: Improving DARTS by Node Normalization and Decorrelation Discretization
EvNAS-C74.9-4.925.1Evolving Neural Architecture Using One Shot Model
SharpSepConvDARTS74.1-4.9M25.1sharpDARTS: Faster and More Accurate Differentiable Architecture Search
FairNAS-B75.1-4.5M24.9FairNAS: Rethinking Evaluation Fairness of Weight Sharing Neural Architecture Search
ProxylesNAS75.1-5.1M24.9ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware
iDARTS (ImageNet)--5.1M24.7iDARTS: Improving DARTS by Node Normalization and Decorrelation Discretization
0 of 135 row(s) selected.