Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Apr 1, 2026

• Add SmolLM2 smol-smoltalk evaluation pipeline and config (#409)

  Show description for 8cf1ccc
  baochunli

  authored
  8cf1ccc

  Copy full SHA for 8cf1ccc

Commits on Mar 10, 2026

• Added SmolVLA and LeRobot integration with configs, validation, and documentation. (#407)

  Show description for da88695
  baochunli

  authored
  da88695

  Copy full SHA for da88695

Commits on Jan 9, 2026

• Added Benchmark Evaluation Framework with CORE Benchmark Suite (#403)

  Show description for 86e22d5
  

  3 peopleauthored
  86e22d5

  Copy full SHA for 86e22d5

Commits on Oct 22, 2025

• Added Torchvision as a new datasource, replacing MNIST, EMNIST, FashionMNIST, CelebA, CIFAR10, and CIFAR100. (#399)

  baochunli

  authored
  e988a96

  Copy full SHA for e988a96

• Added MLX as an alternative backend to PyTorch; changed file format for model saving to .safetensors (supported by both MLX and PyTorch). (#398)

  Show description for f938fad
  baochunli

  authored
  f938fad

  Copy full SHA for f938fad

Commits on Oct 20, 2025

• Cleaned up unused datasources, configuration files, and base classes.

  baochunli

  committed
  83cc1f7

  Copy full SHA for 83cc1f7

• Migrated the entire project from YAML to TOML as the configuration file format. (#397)

  Show description for 61b9dd5
  baochunli

  authored
  61b9dd5

  Copy full SHA for 61b9dd5

Commits on Oct 17, 2025

• Migrated all examples off legacy client API; removed all legacy client API support for old subclassing and hooks; redesigned and strengthened tests; refactored HuggingFace trainer and included LoRA…

  Show description for d4bcf7e
  baochunli

  authored
  d4bcf7e

  Copy full SHA for d4bcf7e

Commits on Oct 16, 2025

• Add an multi-party computation runtime for additive and Shamir support. (#392)

  Show description for a43967f
  baochunli

  authored
  a43967f

  Copy full SHA for a43967f

• Redesigned the client API to using composition-based strategies. (#391)

  baochunli

  authored
  06fec04

  Copy full SHA for 06fec04

Commits on Oct 13, 2025

• Redesigned the trainer API to use the strategy pattern for dependency injection, rather than inheritance and hooks. (#388)

  baochunli

  authored
  20380c4

  Copy full SHA for 20380c4

Commits on Oct 11, 2025

• Refactored the entire documentation website and all the examples. (#385)

  Show description for f96269a
  

  Jasmine-Yuting-Zhang

  and

  baochunli

  authored
  f96269a

  Copy full SHA for f96269a

Commits on Sep 30, 2025

• Migrated the Python package manager to uv, and retired MindSpore/TensorFlow support. (#383)

  Show description for f2134eb
  

  Jasmine-Yuting-Zhang

  and

  baochunli

  authored
  f2134eb

  Copy full SHA for f2134eb

Commits on Nov 5, 2023

• Fixed an issue where Federated EMNIST breaks with client side testing.

  baochunli

  committed
  e070888

  Copy full SHA for e070888

Commits on Oct 21, 2023

• Merged examples/split_learning into Plato code base and added some callbacks functions as APIs for split learning. (#357)

  Show description for d37ddab
  

  dixiyao

  and

  baochunli

  authored
  d37ddab

  Copy full SHA for d37ddab

Commits on Oct 8, 2023

• Removed unnecessary configuration files for YOLOv8.

  baochunli

  committed
  3d927f9

  Copy full SHA for 3d927f9

• Removed unnecessary YOLOv8 configuration files.

  baochunli

  committed
  453bb1d

  Copy full SHA for 453bb1d

Commits on Oct 7, 2023

• Added YOLOv8 COCO128 config files.

  baochunli

  committed
  d50fa56

  Copy full SHA for d50fa56

• Removed YOLOv5 support.

  baochunli

  committed
  a0c5707

  Copy full SHA for a0c5707

• Added preliminary YOLOv8 support.

  baochunli

  committed
  02b25ae

  Copy full SHA for 02b25ae

Commits on Apr 19, 2023

• Reduced the scale of the fedavg_async_lenet5 configuration to mitigate CI Test issues.

  baochunli

  committed
  fb6f50e

  Copy full SHA for fb6f50e

• Fixed an issue in the distribution non-iid sampler. Closes #308.

  baochunli

  committed
  1bfa1e5

  Copy full SHA for 1bfa1e5

Commits on Mar 29, 2023

• renamed the module name and moved the place of importing module. (#302)

  dixiyao

  authored
  bbe2950

  Copy full SHA for bbe2950

• Revert "renamed the name in the import module."

  Show description for 805a6e0
  dixiyao

  committed
  805a6e0

  Copy full SHA for 805a6e0

• renamed the name in the import module.

  dixiyao

  committed
  cb09dea

  Copy full SHA for cb09dea

Commits on Jan 28, 2023

• Femnist bug (#292)

  dixiyao

  authored
  8c0ec91

  Copy full SHA for 8c0ec91

Commits on Dec 4, 2022

• Improved the handling of failed clients on the server. Closes #176. (#266)

  Show description for a878536
  

  3 peopleauthored
  a878536

  Copy full SHA for a878536

Commits on Nov 28, 2022

• Implemented a general Multi-layer Perceptron (MLP) model generator. (#267)

  Show description for 39d709a
  

  CSJDeveloper

  and

  baochunli

  authored
  39d709a

  Copy full SHA for 39d709a

Commits on Nov 21, 2022

• Allowed ViT models from HuggingFace to be trained from scratch; Added T2T-ViT and DeepViT models. (#261)

  Show description for 8f35014
  

  dixiyao

  and

  baochunli

  authored
  8f35014

  Copy full SHA for 8f35014

Commits on Oct 28, 2022

• Added HuggingFace ViT into the list of supported model types. (#254)

  Show description for 492a6ce
  

  dixiyao

  and

  baochunli

  authored
  492a6ce

  Copy full SHA for 492a6ce

• Implemented processors and server with homomorphic encryption support as well as an example. (#252)

  Show description for fa36e6c
  

  HeyHao

  and

  baochunli

  authored
  fa36e6c

  Copy full SHA for fa36e6c

Commits on Oct 27, 2022

• Fixed an issue that prevented the timm scheduler to work correctly when global_lr_scheduler is true. (#253)

  Show description for 261fd61
  

  dixiyao

  and

  baochunli

  authored
  261fd61

  Copy full SHA for 261fd61

Commits on Aug 20, 2022

• Removed plot.py and its configuration; Refactored FedAtt; Improved examples documentation; Released 0.4.4.

  baochunli

  committed
  b0a11dd

  Copy full SHA for b0a11dd

Commits on Aug 15, 2022

• Supported a run history of per-epoch training loss values and other metrics in the basic trainer; added support for training with LR schedulers from timm. (#231)

  Show description for 5136c61
  baochunli

  authored
  5136c61

  Copy full SHA for 5136c61

• Renamed loss function to loss criterion; Updated the documentation for configuration settings.

  baochunli

  committed
  aaea954

  Copy full SHA for aaea954