conftrace_

activation checkpointing

4 papers

Explore in graph

Co-occurring keywords · papers mentioning both

memory optimization (3) deep learning (2) layer replacement (1) eviction policy (1) memory budget (1) transformer model (1) tensor rematerialization (1) computational graph (1) gradient accumulation (1) computation graph (1)

Papers

Rockmate: an Efficient, Fast, Automatic and Generic Tool for Re-materialization in PyTorch ICML 2023 Tempo: Accelerating Transformer-Based Model Training through Memory Footprint Reduction NeurIPS 2022 Dynamic Tensor Rematerialization ICLR 2021 Efficient Rematerialization for Deep Networks NeurIPS 2019