Function Approximation on 1D approximation benchmark (test)
1.05Test MSEFourier Features
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Fourier Featureswidth=50, depth=6, scale σ=128, #modes M=32, #trainable-parameters=16 0512025.02 | 1.05 | 2.43 | |
| Fourier Featureswidth=37, depth=6, scale σ=2, #modes M=128, #trainable-parameters=16 5772025.02 | 1.05 | 1.93 | |
| Fourier Featureswidth=45, depth=6, scale σ=64, #modes M=64, #trainable-parameters=16 2012025.02 | 1.06 | 8.14 | |
| Fourier Featureswidth=37, depth=6, scale σ=16, #modes M=128, #trainable-parameters=16 5772025.02 | 1.1 | 7.72 | |
| Fourier Featureswidth=45, depth=6, scale σ=16, #modes M=64, #trainable-parameters=16 2012025.02 | 1.14 | 8.32 | |
| Fourier Featureswidth=50, depth=6, scale σ=32, #modes M=32, #trainable-parameters=16 0512025.02 | 1.16 | 8.34 | |
| Fourier Featureswidth=37, depth=6, scale σ=4, #modes M=128, #trainable-parameters=16 5772025.02 | 1.16 | 2.63 | |
| Fourier Featureswidth=50, depth=6, scale σ=16, #modes M=32, #trainable-parameters=16 0512025.02 | 1.24 | 8.4 | |
| Fourier Featureswidth=45, depth=6, scale σ=8, #modes M=64, #trainable-parameters=16 2012025.02 | 1.36 | 8.19 | |
| Fourier Featureswidth=50, depth=6, scale σ=1, #modes M=32, #trainable-parameters=16 0512025.02 | 1.38 | 6.41 | |
| Fourier Featureswidth=37, depth=6, scale σ=8, #modes M=128, #trainable-parameters=16 5772025.02 | 1.6 | 9.7 | |
| Fourier Featureswidth=37, depth=6, scale σ=64, #modes M=128, #trainable-parameters=16 5772025.02 | 1.61 | 8.06 | |
| Fourier Featureswidth=37, depth=6, scale σ=128, #modes M=128, #trainable-parameters=16 5772025.02 | 1.67 | 1.06 | |
| Fourier Featureswidth=50, depth=6, scale σ=8, #modes M=32, #trainable-parameters=16 0512025.02 | 1.69 | 9.5 | |
| Fourier Featureswidth=50, depth=6, scale σ=64, #modes M=32, #trainable-parameters=16 0512025.02 | 1.72 | 1.05 | |
| Fourier Featureswidth=45, depth=6, scale σ=1, #modes M=64, #trainable-parameters=16 2012025.02 | 1.89 | 7.43 | |
| FMMNNwidth=200, rank=16, depth=6, #trainable-parameters=16 2812025.02 | 2.77 | 2.41 | |
| Fourier Featureswidth=50, depth=6, scale σ=2, #modes M=32, #trainable-parameters=16 0512025.02 | 3.12 | 9.31 | |
| Fourier Featureswidth=45, depth=6, scale σ=128, #modes M=64, #trainable-parameters=16 2012025.02 | 3.45 | 1.41 | |
| Fourier Featureswidth=37, depth=6, scale σ=1, #modes M=128, #trainable-parameters=16 5772025.02 | 4.12 | 9.75 | |
| Fourier Featureswidth=50, depth=6, scale σ=4, #modes M=32, #trainable-parameters=16 0512025.02 | 4.56 | 1.66 | |
| Fourier Featureswidth=45, depth=6, scale σ=4, #modes M=64, #trainable-parameters=16 2012025.02 | 5.15 | 1.81 | |
| Fourier Featureswidth=45, depth=6, scale σ=2, #modes M=64, #trainable-parameters=16 2012025.02 | 7.45 | 5.95 | |
| Fourier Featureswidth=37, depth=6, scale σ=32, #modes M=128, #trainable-parameters=16 5772025.02 | 8.68 | 6.84 | |
| Fourier Featureswidth=45, depth=6, scale σ=32, #modes M=64, #trainable-parameters=16 2012025.02 | 8.81 | 6.97 |