2D Synthetic Target Sampling on Gaussian 2D Synthetic
0.013KSDSVGD(500)
Evaluation Results
| Method | Links | |
|---|---|---|
| SVGD(500)Sampler Category=MCMC, Stepsize=0.01, Iterations=500, num particles=500, num chains=202023.06 | 0.013 | |
| STEINGANSampler Category=NEURAL, Architecture=4-layer MLP with 400 hidden units, Activation=ELU2023.06 | 0.091 | |
| HMC(500)Sampler Category=MCMC, Stepsize=0.1, Iterations=500, num particles=500, num chains=202023.06 | 0.094 | |
| FISHER-NSSampler Category=NEURAL, Architecture=4-layer MLP with 400 hidden units, Activation=ELU, Score Network updates=52023.06 | 0.095 | |
| KL-NSSampler Category=NEURAL, Architecture=4-layer MLP with 400 hidden units, Activation=ELU, Score Network updates=52023.06 | 0.099 | |
| COUP-FLOWSampler Category=NEURAL, Blocks=3, Architecture=4-layer MLP with 200 hidden units, Activation=GELU2023.06 | 0.102 | |
| LD(500)Sampler Category=MCMC, Stepsize=0.01, Iterations=500, num particles=500, num chains=202023.06 | 0.107 | |
| KSD-NSSampler Category=NEURAL, Architecture=4-layer MLP with 400 hidden units, Activation=ELU2023.06 | 0.206 |