List your company

Finding Optimal Elimination Order on Transformer

4,656Number of MultiplicationsAlphaGrad
-558.1634,637.4269,833105,028.58Jun 7, 2024

Evaluation results

MethodLinks
AlphaGradMCTS simulations=250, Batch size=1, Gumbel noise scale=1, Reward scaling=log-scaling2024.064,656
ReverseBatch size=12024.064,688
AlphaGradMCTS simulations=50, Batch size=1, Gumbel noise scale=1, Reward scaling=log-scaling2024.064,831
MarkowitzBatch size=12024.0651,869
ForwardBatch size=12024.06135,010