2.7.1: official release with feature improvements and bug fixes

2.7.1 Release Contributors (Chronological Order)

The following lists all contributors to the NVFLARE 2.7.1 release (from 2.7.0 to 2.7.1), ordered by when they first contributed.

Total commits in this release: 60
nvkevlu - 11 commits
Yuan-Ting Hsieh (謝沅廷) - 11 commits
Chester Chen - 6 commits
dependabot[bot] - 2 commits
Holger Roth - 6 commits
Yuanyuan Chen - 5 commits
Isaac Yang - 10 commits
Nick - 1 commit
Ruda Porto Filgueiras - 1 commit
Zhihong Zhang - 4 commits
Ziyue Xu - 3 commits

Welcome First-Time Contributors!

We would like to extend a special thank you to the following contributors who made their first commits to NVFLARE in this release:

Nick (@nickl1234567) - 1 commit - PR #3806
Ruda Porto Filgueiras (@r.filgueiras) - 1 commit - PR #3741

Thank you for your contributions to the NVFLARE community!

What's Changed

• Fix webpage links by @nvkevlu in #3812
• Bump vite from 6.3.6 to 6.4.1 in /web by @dependabot[bot] in #3807
• Update Hello-PT and Stats by @holgerroth in #3810
• Bump astro from 5.13.2 to 5.14.4 in /web by @dependabot[bot] in #3776
• Use Python 3.9 typing by @cyyever in #3611
• Fix docs and add missing diagram by @nvkevlu in #3815
• Simplify quantization code by @cyyever in #3612
• Adjust supported minimum Python versions to 3.9 by @cyyever in #3665
• Enhance MLflow receiver by @YuanTingHsieh in #3657
• Fix edge simulator by @YuanTingHsieh in #3813
• Deployment guide of confidential ACI [skip ci] by @IsaacYangSLA in #3816
• Add Azure CVM deployment guide [skip ci] by @IsaacYangSLA in #3820
• Update release notes by @YuanTingHsieh in #3818
• Fix a syntax error in ACI doc [skip ci] by @IsaacYangSLA in #3823
• Update sub_start.sh with one additional option by @IsaacYangSLA in #3821
• Update cc docs by @YuanTingHsieh in #3824
• Add FedNCA publication by @nickl1234567 in #3806
• CC document updates and others by @IsaacYangSLA in #3826
• Add Tensor Stream component for efficient safetensors-based model tensor streaming by @rfilgueiras in #3741
• Add GPU CC docs by @YuanTingHsieh in #3825
• Update FAQ by @nvkevlu in #3832
• Android app enhancements for state management and status on UI by @nvkevlu in #3819
• New CC token verification mechanism by @YuanTingHsieh in #3829
• Add ADAQUANT quantization scheme by @cyyever in #3628
• Add CC GPU notes by @YuanTingHsieh in #3840
• Update cc doc structure by @YuanTingHsieh in #3843
• Added User Data Instrucitons by @nvidianz in #3844
• Fix for quantization by @ZiyueXu77 in #3846
• Added unmount instructions by @nvidianz in #3847
• Fix cc provision example by @YuanTingHsieh in #3845
• Update document not to include packager in the Azure project.yml [skip ci] by @IsaacYangSLA in #3851
• Update links by @YuanTingHsieh in #3852
• [HuggingFace] Multi-node client support by @holgerroth in #3755
• Multi-node training: Add wrapper to update master address in client_api_config.json by @holgerroth in #3855
• Fix format issues of some yaml in the document [skip ci] by @IsaacYangSLA in #3857
• Fix broken links to some images by @nvkevlu in #3854
• Fix hf llm SimEnv call by @holgerroth in #3859
• Fix Android edge device stability and resume support by @nvkevlu in #3858
• Add SklearnFedAvgRecipe and convert sklearn-linear example to recipe format by @nvkevlu in #3835
• Add XGBBaggingRecipe and convert random_forest example to recipe format by @nvkevlu in #3838
• Add SVMFedAvgRecipe and convert sklearn-svm example to recipe format by @nvkevlu in #3837
• Add KMeansFedAvgRecipe and convert sklearn-kmeans example to recipe by @nvkevlu in #3836
• Fix a literalinclude link by @IsaacYangSLA in #3860
• Add check on empty cc_info by @IsaacYangSLA in #3862
• Fix broken quantization pipeline by recent adaquant by @ZiyueXu77 in #3861
• Use absolute path for nvidia logo [skip ci] by @IsaacYangSLA in #3866
• Address race issue for quantization filter by @ZiyueXu77 in #3865
• Fixed cc_provision example by @nvidianz in #3864
• [LLM Tutorial] Restructure codes by @holgerroth in #3850
• [NVFlare Summer of Code] Adding FedOBD (IJCAI 2023) to the research dir by @cyyever in #3666
• LLM tutorial streaming example update by @holgerroth in #3868
• Moved several messages to debug level by @nvidianz in #3869

Full Changelog: 2.7.0...2.7.1

Feature highlights

Tensor-Streaming Example

https://github.com/NVIDIA/NVFlare/tree/main/examples/advanced/tensor-stream

FedNCA - Equitable Federated Learning with NCA
https://github.com/NVIDIA/NVFlare/tree/main/research/FedNCA

FedOBD: Opportunistic Block Dropout for Efficiently Training Large-scale Neural Networks through Federated Learning
https://github.com/NVIDIA/NVFlare/tree/main/research/fedobd

Multi-Node Training Solution
https://github.com/NVIDIA/NVFlare/blob/main/examples/advanced/llm_hf/MULTINODE.md