NVIDIA FLARE Files

What's Changed

• [2.7] Address client-side vulneralbility for tree-based horizontal XGB by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3896
• [2.7] Allow one project admin by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/3897
• [2.7] Cherry pick fixes from main by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/3910
• [2.7] Change default scheme to HTTP by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/3909
• [2.7] Downloader Rework Backport by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/3908
• [2.7] Remove overseer codes by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/3899
• [2.7] Lower Downloader logging verbosity by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/3913
• [2.7] Cherry pick Convert PSI example to use recipe (#3901) by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/3911
• [2.7] Redesign Cifar10 PT by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3905
• [2.7] Restructure CIFAR-10 PT example by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3916
• [2.7] Refactor CIFAR-10 TensorFlow example by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3919
• [2.7] Consolidate BaseFedJob and fedavg.py and Fix import error for TBAnalyticsReceiver by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3918
• [2.7] Cherry pick from main by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/3920
• [2.7] Cherry pick of Fix preflight check and ci (#3917) by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/3929
• [2.7] Fix Job API TF examples by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3926
• [2.7] Kaplan Meier updates by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/3935
• [2.7] Multinode doc update by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3939
• [2.7] Add missing tensorboard requirements by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3938
• [2.7] Cherry pick cross site eval 3923 by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3936
• [2.7] Improve error message in client API by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3947
• [2.7] Removed references to ws by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/3950
• [2.7] Cherry pick 3930 3945 by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/3952
• [2.7] Cherry-pick of Consolidate LR examples (#3944) by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/3953
• [2.7] Expose launch_once option in ScriptRunner by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3954
• [2.7] Cherry pick simplify CSE recipe (#3942) by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3957
• [2.7] Cifar10 tf central training logs path by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3955
• [2.7] BioNeMo Recipes by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3943
• [2.7] Apply same llm updates to 2.7 by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/3956
• [2.7] Apply same GNN updates to 2.7 by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/3966
• [2.7] cherry pick fix for NumpyFedAvgRecipe experiment tracking by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3970
• [2.7] Remove step-by-step by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/3972
• [2.7] Hello Differential Privacy by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3961
• [2.7] expose key metric in recipe by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/3973
• [2.7] Update key_metric over all examples by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/3981
• [2.7] Raise exception on FOBS errors by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/3969
• [2.7] Fix TLS corruption by replacing fork with posix_spawn (#3856) by @GeorgeWang-nv in https://github.com/NVIDIA/NVFlare/pull/3983
• [2.7] Temporally remove pre install tool and example from 2.7 branch by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/3986
• [2.7] multinode guide for llm_hf by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/3991
• [2.7] Convert MONAI examples to Recipe by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3971
• [2.7] Cherrypick from main: Fix docs [skip ci] (#3974) by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/3992
• [2.7] BioNeMo Task Fitting with PyTorch by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/3982
• [2.7] Cherry Pick Comprehensively remove mention of SAG by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3997
• [2.7] Cherry pick Make updates to Client API tutorials by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3995
• [2.7] Update web page and bump dependencies for security fixes by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3999
• [2.7] Cherry pick Add recipe for xgboost by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3994
• [2.7] Ignore downloder no ref_id errors by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4004
• [2.7] Cherry Pick Fix hello-numpy-cross-val example by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/3998
• [2.7] FedAvg Merge with FedAvgEarlyStopping + InTimeAggregation by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/3993
• [2.7] Mandating signature by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4008
• [2.7] Fix NumPy CSE regression in 2.7.2rc4 by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4011
• [2.7] handle non exists file with absolute file path by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4007
• [2.7] Cherry pick Add Recipe for Experiment Tracking by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4009
• [2.7] Release news and feature highlights, doc restructure [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4017
• [2.7] Update LR related examples by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4014
• [2.7] Update product feature docs [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4018
• [2.7] Fix sklearn examples by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4015
• [2.7] Increase link check timeout by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4027
• [2.7] Convert AMPLIFY example to recipe by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4022
• [2.7] Recipe site configuration [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4021
• [2.7] Add Brats to research by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4026
• [2.7] Fixed the swarm bug by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4025
• [2.7] Add recipes to init by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4031
• [2.7] Enhance weighted agg helper by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4020
• [2.7] Cherry pick 4002 4005 by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4034
• [2.7] Added origin to the stream lookup key by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4033
• [2.7] Add Server-Side Memory Management for Long-Running FL Jobs by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4042
• [2.7] same switch to open model for llm by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4044
• [2.7] Support bf16 in PT model persistor by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4045
• [2.7] remove no longer supported arg, fix data format by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4047
• [2.7] Add integration tests for examples (cherrypick from [#4041]) by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4051
• [2.7] Move monai examples under advanced by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4053
• [2.7] Update custom authentication example by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4055
• Fixed keycloak docker tag by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4058
• [2.7] increase link check timout by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4060
• [2.7] Increase BioNeMo external script init timeout by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4057
• [2.7] Update Edge for Android by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4064
• [2.7] update GNN readme by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4065
• [2.7] Update info logging of Cacheable by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4062
• [2.7] Fix the rest of the examples by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4039
• [2.7] Updates to notebooks by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4066
• [2.7] Add timeout check by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4070
• [2.7] Update to standardize all cifar10 data location in self-paced tutorails by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4072
• [2.7] Fix/notebooks by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4068
• [2.7] fix skmeans and vertical learning notebooks by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4076
• [2.7] Replace NLP-NER with link to tutorial by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4077
• [2.7] Add missing server_memory_gc_rounds parameter to recipes by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4081
• [2.7] Redesign Job-level Authorization Example by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4074
• [2.7] amplify tutorial updates by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4089
• [2.7] Remove link check on github by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4104
• [2.7] Logging tutorial fix by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4098
• [2.7] Job CLI Tutorial Fixes by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4096
• [2.7] Add dynamic ignore_result_error logic and POC environment cleanup by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4084
• [2.7] Update docker example by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4102
• [2.7] Chapter 1 TensorBoard Streaming Fix by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4103
• [2.7] Update holoscan tutorial by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4095
• [2.7] Improve df_stat example by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4108
• [2.7] Chapter 8 Fixes by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4105
• [2.7]: Fix finance example by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4091
• [2.7] Add model config interface with dict-based model input and initial_ckpt support part 1 of 3 by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4082
• [2.7]: Update the streaming example with more details by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4114
• [2.7] Fix xgboost recipes by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4087
• [2.7] Fix federated_policy example by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4090
• [2.7] Remove multi gpu tf 2.7 by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4116
• [2.7] Add dict config and initial_ckpt support to standard recipes (Part 2) by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4117
• [2.7] Address potential data corruption issue with Streamer by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4100
• [2.7] Snapshot task data only by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4126
• [2.7] Chapter 2 Fixes by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4115
• [2.7] Update jobapi pt example by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4113
• [2.7] Use dict-based config in HuggingFace example by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4123
• [2.7] Chapter 10 Fixes by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4119
• [2.7] improve tensor streaming timeout by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4120
• [2.7] Chapter 12 Fixes by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4122
• [2.7] CIFAR-10 Update tb event reader and requirements [skip ci] by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4125
• [2.7] Updates for latest api from PEFT/TRL by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4128
• [2.7] Add tbparse license [skip ci] by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4136
• [2.7] Touchups on local training scripts by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4137
• [2.7]: Fix the monitoring example by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4099
• [2.7] Cherry-pick LLM HF updates by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4140
• [2.7] Fix POC Run result caching and cleanup by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4132
• [2.7] Recipe Interface Part 3: Dict Model Config and Initial Checkpoint Support by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4130
• [2.7] Fix xgboost adaptor by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4127
• [2.7] Recipe Interface Part 3: Documentation Updates [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4131
• [2.7] Tutorials disclaimer by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4142
• [2.7] Fix swarm controller + tensor streaming issue by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4141
• [2.7] Add comprehensive timeout documentation [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4083
• [2.7] Rename recipe argument 'initial_model' to 'model' by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4144
• [2.7] Fix arg docstring [skip ci] by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4145
• [2.7] added comment in job_recipe notebook by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4147
• [2.7] CIFAR-10 Experiment Tracking Instructions Corrections by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4148
• [2.7] Recipe: support relative initial_ckpt and API enhancements by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4155
• [2.7] Revert model back to initial model for Job API by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4153
• [2.7] Fix fed eval example by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4152
• [2.7] Rename sklearn recipe args by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4154
• [2.7] Recipe API: use class_path for model dict, rename validate_ckpt by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4156
• [2.7] Update links on web page by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4160
• [2.7] Restructure documentation for persona-driven navigation and real-world use cases [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4159
• [2.7] Add retry mechanism for streaming download on TIMEOUT by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4167
• [2.7] Fix stream pool starvation issue by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4171
• [2.7] Fix hello-numpy-cross-val example by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4168
• [2.7] Add end-to-end download starvation test for stream pool fix by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4172
• [2.7] Fix numpy cross val path of results by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4173
• [2.7] update class_path for llm example by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4175
• [2.7] Remove roadmap from documentation [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4179
• [2.7] Initial checkpoint info by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4178
• [2.7] Fix cifar10 integration tests by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4150
• [2.7] Avoid self-message deadlock for local swarm result submission by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4186
• [2.7] smaller lock in produce item by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4174
• [2.7] Replace job.to with alternative recipe apis by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4183
• [2.7] Fix recipes and job api model/initial_model confusion by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4188
• [2.7] Fix numpy cross val sticky property by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4181
• [2.7] Use Initial Global Model in BioNeMo Recipe Examples by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4189
• [2.7] Add PR-4172 style tests for swarm self-result submission fix by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4190
• [2.7] update to match fix in main by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4202
• [2.7] Fix RxTask self-deadlock on stream error cleanup by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4204
• [2.7] Mitigate F3 streaming Head-of-Line (HOL) stalls and add guardrails by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4206
• [2.7] Client-side memory management by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4211
• [2.7] Pass-Through: Zero Tensor Copy at CJ for Large-Model Federated Training by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4210
• [2.7] pin fastdigest==0.4.0 due to API changes by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4217
• [2.7] Skip unsupported metrics for aggregation by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4223
• [2.7] Fix global model selection by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4222
• [2.7] Pin pandas<3.0 and fix pandas 3.x compatibility in federated statistics by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4227
• [2.7] Fix recipe API bug list and harden recipe behavior by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4228
• [2.7] Fix subprocess converter wiring, swarm learning bugs, and recipe enhancements by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4225
• [2.7]: Fix a security issue on FileRetriever by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4230
• [2.7] Fix hierarchical FL startup failures: deployment timeouts, selective client exclusion, and dead-detection debounce by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4209
• [2.7] Fix client-side RSS memory growth and subprocess logging gap by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4231
• [2.7] Update 2.7.2 release notes: streaming hardening, memory management, hierarchical startup stability [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4218
• [2.7] Fix FedOpt argument definitions by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4235
• [2.7] Accept 'class_path' in component config for job API consistency by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4239
• Class path component builder enhancements by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4242
• [2.7] Check full Apache license header and normalize inconsistent headers by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4243
• [2.7] Fix 1-18: large-model subprocess memory, reverse/forward PASS_THROUGH, download gating, MSG_ROOT_TTL, LazyRef local-aggr, logging, min_clients, diagnostics by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4247
• [2.7] Add missing submit_model executor by @nvkevlu in https://github.com/NVIDIA/NVFlare/pull/4254
• [2.7] Fix XGB cyclic by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4255
• [2.7] Fix SimEnv to be compatible with strict simulator_run API by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4250
• [2.7] Skip unsupported metrics for aggregation (#4223) by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4252
• [2.7] Fix swarm recipe model issue by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4260
• [2.7] Fedbuff doc and check updates by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4262
• [2.7] Guard Flower integration against incompatible flwr CLI versions by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4253
• [2.7] Clarify TensorFlow recipe docs to require subclassed Keras models [docs] by @pcnudde in https://github.com/NVIDIA/NVFlare/pull/4264
• [2.7] Improve FedAvg HE TenSEAL context provisioning guidance and validation by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4258
• [2.7] Improve et import by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4259
• [2.7] update example readme [skip ci] by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4272
• [2.7] Fix RC12 Swarm ext-process bugs: msg_root race, CSE model load, validate hang, round_timeout by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4270
• [2.7] add device_wait_timeout to ET class by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4273
• [2.7] Enhancements on CI install_requirements by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4249
• [2.7] Scaffold update: defensive check; docstring update by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4279
• [2.7] 2.7.2 documentation updates [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4278
• [2.7] Fix external kill before server download by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4275
• [2.7] Added fix for RCE bug caused by not validating type_name by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4295
• Add missing types for FOBS by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4298
• [2.7] Fix FilePipe TOCTOU by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4296
• [2.7] Fix pt_init_client data_path to avoid concurrent CIFAR10 download corruption by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4297
• [2.7] Datasets license; swarm readme update [skip ci] by @holgerroth in https://github.com/NVIDIA/NVFlare/pull/4303
• [2.7] Docs update to include CVM base image building instructions by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4306
• [2.7]: Sign the root folders of startup kits by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4301
• [2.7]: Fix an inconsistency between document and codes by @IsaacYangSLA in https://github.com/NVIDIA/NVFlare/pull/4308
• [2.7] Receiver-side PASS_THROUGH for ext-process CJ forward path by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4309
• [2.7] increase FilePipe default heartbeat timeout value to 600s by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4307
• [2.7] Fix receiver-side PASS_THROUGH channel mismatch for GET_TASK by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4312
• [2.7] Fix auth test by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4315
• [2.7] Fix Swarm deadlock: _executing guard prevents pipe handler replacement mid-transaction by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4314
• [2.7] Condense 2.7.2 release notes to high-level feature highlights [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4316
• [2.7] Fix SignatureBuilder: sign root only in CC mode, revert to startup+local for non-CC by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4318
• [2.7] Fix FedAvg custom aggregator losing fl_ctx CURRENT_ROUND by @ZiyueXu77 in https://github.com/NVIDIA/NVFlare/pull/4317
• [2.7] Update 2.7.2 release notes with CWE IDs for security fixes [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4319
• [2.7] Polish 2.7.2 release notes memory section and add benchmark tables [skip-ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4323
• [2.7] Clear pipe state between rounds to prevent stale PEER_GONE by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4326
• [2.7] Redact security fix implementation details from release notes [skip ci] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4332
• [2.7] Fix MetricRelay multi-round pipe lifecycle and bad-payload crash by @YuanTingHsieh in https://github.com/NVIDIA/NVFlare/pull/4334
• [2.7] Added monai type to white-list by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4320
• [2.7] Add FOBS security section to decomposer_for_large_object.rst [skip ci] by @nvidianz in https://github.com/NVIDIA/NVFlare/pull/4325
• [2.7] Smooth memory benchmark summary paragraph in 2.7.2 release notes [skip cli] by @chesterxgchen in https://github.com/NVIDIA/NVFlare/pull/4337

Full Changelog: https://github.com/NVIDIA/NVFlare/compare/2.7.1...2.7.2