torch-sparse-adamw

No project description provided

Project description

# Pytorch Sparse AdamW This repository contains the sparse version of AdamW optimizer.

The SparseAdamW optimizer behaves like AdamW optimizer, but updates only the statistics for gradients which are computed, in the same way as SparseAdam optimizer. The optimizer can only be used on modules, which produce sparse gradients, e.g., nn.Embedding.

## Install Install by running: `bash pip install torch-sparse-adamw `

Project details

Release history Release notifications | RSS feed

This version

1.0.0

Jan 25, 2021

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

torch_sparse_adamw-1.0.0.tar.gz (4.9 kB view hashes)

Uploaded Jan 25, 2021 Source

Built Distribution

torch_sparse_adamw-1.0.0-py3-none-any.whl (4.6 kB view hashes)

Uploaded Jan 25, 2021 Python 3

Hashes for torch_sparse_adamw-1.0.0.tar.gz

Hashes for torch_sparse_adamw-1.0.0.tar.gz
Algorithm	Hash digest
SHA256	`33358fe69a9624f102d34cb6902ac01632162e4891bd83d973dfffed0c2c934e`
MD5	`3e6f39a879de442be90657d4bbaf11bc`
BLAKE2b-256	`c951ca6570718f4ce1339d53422131121ad048e1a699256159f061fa18368784`

Hashes for torch_sparse_adamw-1.0.0-py3-none-any.whl

Hashes for torch_sparse_adamw-1.0.0-py3-none-any.whl
Algorithm	Hash digest
SHA256	`0729195d888716ecdfd45cfc7f6d807589c0e4942e4a8256c36b9b4c49b4fba1`
MD5	`5aa7f3936f3618292cf5c9097b9a2eda`
BLAKE2b-256	`d99f1811265caceb27ffe317d30e7e033fcc9679cfa34966ad5b970e4fd382b9`