0/1 Knapsack Problem in PyTorch

← main page pytorch algorithms compilers optimization Use 20x less peak RAM with dp_knapsack_sliding_hirschberg, a new activation memory budget solver for PyTorch TL;DR: If you use a default torch._functorch.config.activation_memory_budget_solver=dp, you can change it to dp_knapsack_sliding_hirschberg and decrease your peak RAM usage during memory planning by a factor of 20. What? When PyTorch builds a computation graph of your ML model, it tries many different things to make it fast and memory