Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on May 8, 2024

• Revert "successfully run locally, and modify config.yaml(CIFAR10)"

  Show description for 1f644e9
  zhutmost

  committed
  1f644e9

  Copy full SHA for 1f644e9

• Revert "fix util/data_loader.py __balance_val_split & util/lr_scheduler.py ExponentialLr bugs"

  Show description for 769a129
  zhutmost

  committed
  769a129

  Copy full SHA for 769a129

• fix util/data_loader.py __balance_val_split & util/lr_scheduler.py ExponentialLr bugs

  

  jachin

  authored and

  zhutmost

  committed
  1d739ec

  Copy full SHA for 1d739ec

• successfully run locally, and modify config.yaml(CIFAR10)

  

  Jachin Jiang

  authored and

  zhutmost

  committed
  f478eec

  Copy full SHA for f478eec

Commits on Jul 23, 2020

• feat: release an example model

  zhutmost

  committed
  2c24a96

  Copy full SHA for 2c24a96

Commits on Jul 18, 2020

• feat: dump the model graph in TensorBoard

  zhutmost

  committed
  9cade11

  Copy full SHA for 9cade11

Commits on Jul 17, 2020

• style: format the code with PEP8

  zhutmost

  committed
  00673a2

  Copy full SHA for 00673a2

• fix: fix the logging.conf location

  zhutmost

  committed
  d902d3f

  Copy full SHA for d902d3f

Commits on Jun 15, 2020

• feat: add ResNet for CIFAR10 dataset

  zhutmost

  committed
  aafa974

  Copy full SHA for aafa974

Commits on Jun 7, 2020

• fix: move batch_size to dataloader configurations

  zhutmost

  committed
  ff428ea

  Copy full SHA for ff428ea

• fix: fix grad_scale in channel-wise quantization

  zhutmost

  committed
  5b43c67

  Copy full SHA for 5b43c67

• fix: allow to load datasets in a deterministic way

  zhutmost

  committed
  9b642de

  Copy full SHA for 9b642de

Commits on Jun 6, 2020

• feat: allow to quantize a tensor along its output channel

  zhutmost

  committed
  d5f228a

  Copy full SHA for d5f228a

• fix: split the training and validation sets in balance

  zhutmost

  committed
  02f802e

  Copy full SHA for 02f802e

Commits on Jun 5, 2020

• feat: dump args in a YAML file

  zhutmost

  committed
  cc66547

  Copy full SHA for cc66547

Commits on Jun 1, 2020

• feat: reimplement as non-invasive quantization

  zhutmost

  committed
  c0d319c

  Copy full SHA for c0d319c

Commits on May 31, 2020

• fix: fix a tiny print typo

  zhutmost

  committed
  eb967eb

  Copy full SHA for eb967eb

Commits on May 30, 2020

• feat: remove SwitchableNorm and update README

  zhutmost

  committed
  d708bca

  Copy full SHA for d708bca

• fix: add missing bracket

  zhutmost

  committed
  e95bff4

  Copy full SHA for e95bff4

• feat: replace detach and identity with built-in operators

  zhutmost

  committed
  653dcec

  Copy full SHA for 653dcec

• feat: add CIFAR-10 dataset support

  zhutmost

  committed
  019b895

  Copy full SHA for 019b895

• feat: check if modules specified in excepts exist

  zhutmost

  committed
  a43be4f

  Copy full SHA for a43be4f

Commits on May 26, 2020

• chore: bump pytorch to a new version 1.5

  zhutmost

  committed
  423ef4c

  Copy full SHA for 423ef4c

• feat: allow users to split the training set for cross validation

  zhutmost

  committed
  f2a9b93

  Copy full SHA for f2a9b93

• fix: correct missing parameters in add_module()

  zhutmost

  committed
  77668cd

  Copy full SHA for 77668cd

Commits on May 24, 2020

• fix: correct quantized weight bit width of excepts

  zhutmost

  authored
  9e89fbc

  Copy full SHA for 9e89fbc

Commits on Feb 7, 2020

• docs: prepare for public release

  zhutmost

  committed
  60c2077

  Copy full SHA for 60c2077

Commits on Jan 5, 2020

• fix: correct the config load logic when some keys missed

  zhutmost

  committed
  17ce4a8

  Copy full SHA for 17ce4a8

• feat: simplify the code of PerformanceScoreboard

  zhutmost

  committed
  afea660

  Copy full SHA for afea660

Commits on Jan 4, 2020

• docs: update README file

  zhutmost

  committed
  3b94a7b

  Copy full SHA for 3b94a7b

Commits on Dec 24, 2019

• feat: quantize convolution layers into various bit widths

  zhutmost

  committed
  e9fc672

  Copy full SHA for e9fc672

• feat: log more details in LrScheduler.__str__

  zhutmost

  committed
  c3f276a

  Copy full SHA for c3f276a

• feat: use YAML file to configure instead of command line arguments

  zhutmost

  committed
  8401733

  Copy full SHA for 8401733

Commits on Dec 11, 2019

• fix: raise error when quantizing a conv layer with biases

  zhutmost

  committed
  a6c1f7d

  Copy full SHA for a6c1f7d

• improvement: print more setting details before processing

  zhutmost

  committed
  80e08c1

  Copy full SHA for 80e08c1

Pagination

Previous

Next