Image Classification on DEIC Benchmark

Metric: Average Balanced Accuracy (across datasets) (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	Average Balanced Accuracy (across datasets)▼	Extra Data	Paper	Date↕	Code
1	Harmonic Networks	68.7	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
2	Cross-Entropy baseline	67.9	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
3	Cosine + Cross-Entropy Loss	64.92	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
4	T-vMF Similarity	64.67	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
5	DSK Networks	64.64	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
6	OLÉ	64.15	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
7	Cosine Loss	62.73	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
8	Full Convolution	62.06	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
9	Deep Hybrid Networks	60.33	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code
10	Grad-l2 Penalty	55.47	No	Tune It or Don't Use It: Benchmarking Data-Effic...	2021-08-30	Code

#1Harmonic NetworksSOTA
68.7
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#2Cross-Entropy baseline
67.9
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#3Cosine + Cross-Entropy Loss
64.92
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#4T-vMF Similarity
64.67
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#5DSK Networks
64.64
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#6OLÉ
64.15
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#7Cosine Loss
62.73
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#8Full Convolution
62.06
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#9Deep Hybrid Networks
60.33
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code
#10Grad-l2 Penalty
55.47
Average Balanced Accuracy (across datasets)· 2021-08-30
Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification Code