Stochastic Distributed Optimization
0Convergence RateVR-EXTRA/DIGing
Evaluation results
| Method | Links | |
|---|---|---|
| VR-EXTRA/DIGingvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=|B|, mi every 1/p, assumpt.=n.c.2025.01 | 0 | |
| VR-EXTRA/DIGingvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=|B|, mi every 1/p, assumpt.=s.c.2025.01 | 0 | |
| GT-VRvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=|B|, mi every 1/p, assumpt.=n.c.2025.01 | 0 | |
| GT-VRvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=|B|, mi every 1/p, assumpt.=s.c.2025.01 | 0 | |
| GT-SAGAvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=1, assumpt.=s.c.2025.01 | 0 | |
| GT-SAGAvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=1, assumpt.=n.c.2025.01 | 0 | |
| GT-SARAHvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=|B|, mi every τ, assumpt.=n.c.2025.01 | 0 | |
| GT-SVRGvariance reduction=✓, grad. steps ÷ comm.=1 ÷ 1, # stored variables=3, comm. size=2|Ni|, # ∇fi,j evaluations per iteration=1, mi every τ, assumpt.=s.c.2025.01 | 0 | |
| LT-ADMM-VRvariance reduction=✓, grad. steps ÷ comm.=τ ÷ 1, # stored variables=|Ni| + 1, comm. size=|Ni|, # ∇fi,j evaluations per iteration=|B|, mi every τ, assumpt.=n.c.2025.01 | 0 |