REVIEW 1 cited by
On Efficient Constructions of Checkpoints
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
abstract
Efficient construction of checkpoints/snapshots is a critical tool for training and diagnosing deep learning models. In this paper, we propose a lossy compression scheme for checkpoint constructions (called LC-Checkpoint). LC-Checkpoint simultaneously maximizes the compression rate and optimizes the recovery speed, under the assumption that SGD is used to train the model. LC-Checkpointuses quantization and priority promotion to store the most crucial information for SGD to recover, and then uses a Huffman coding to leverage the non-uniform distribution of the gradient scales. Our extensive experiments show that LC-Checkpoint achieves a compression rate up to $28\times$ and recovery speedup up to $5.77\times$ over a state-of-the-art algorithm (SCAR).
Forward citations
Cited by 1 Pith paper
-
An Efficient Compression of Deep Neural Network Checkpoints Based on Prediction and Context Modeling
Using previous checkpoint values as LSTM context for arithmetic coding reduces compressed checkpoint size by 14% to 31% over ExCP on ViT-L32 and Pythia-410M, with the coding stage lossless.
Discussion (0). Continue with ORCID to comment.