Issues

• [BUG] Validate the tasks argument of job filters and executors (str is split into characters, None crashes)

  NVIDIA/NVFlare#5778 · Filter by author sohailshk opened on Nov 13, 2026

• [BUG] Sending a non-contiguous torch tensor fails in safetensors (e.g. channels_last conv weights)

  NVIDIA/NVFlare#5777 · Filter by author sohailshk opened on Nov 13, 2026

• [BUG] torch.nn.Parameter values cannot be sent with FOBS (receiver rejects the type; whitelisting it gives empty tensors)

  NVIDIA/NVFlare#5776 · Filter by author sohailshk opened on Nov 13, 2026

• Scope Dependabot security updates by example and dependency family

  NVIDIA/NVFlare#5774 · Filter by author holgerroth opened on Nov 13, 2026

• Upgrade Flower integration to the HTTP Fleet API

  NVIDIA/NVFlare#5767 · Filter by author holgerroth opened on Nov 12, 2026

• Support CPU-only Docker client jobs on a Slurm gateway for task execution

  NVIDIA/NVFlare#5755 · Filter by author YuanTingHsieh opened on Nov 10, 2026

• [BUG] @experimental overwrites the caller's warning policy

  bug
  NVIDIA/NVFlare#5753 · Filter by author holgerroth opened on Nov 10, 2026
  

• Add CustomAggregator for reusable custom model aggregation functions

  NVIDIA/NVFlare#5752 · Filter by author holgerroth opened on Nov 10, 2026
  

• Lightning fit-per-round: a round with no optimizer step is sent and FedAvg counts it with weight 1

  NVIDIA/NVFlare#5748 · Filter by author rfilgueiras opened on Nov 10, 2026
  2

• Add a MemCo research example using the high-level Collab API

  NVIDIA/NVFlare#5724 · Filter by author SYannL opened on Nov 6, 2026

• Block-wise INT4 quantization filters for parameter communication

  NVIDIA/NVFlare#5716 · Filter by author AhmadHassan-BTed opened on Nov 6, 2026

• Unify NVFlight builds under nvflare-light and retire nvflare-light-nightly

  NVIDIA/NVFlare#5702 · Filter by author holgerroth opened on Nov 5, 2026 · 2.10

• Step-level model sync for the Lightning Client API

  NVIDIA/NVFlare#5699 · Filter by author rfilgueiras opened on Nov 5, 2026

• [BUG] AsyncNumAssessor aggregates repeated device reports twice

  NVIDIA/NVFlare#5692 · Filter by author 0z5a opened on Nov 4, 2026

• [BUG] AsyncNumAssessor can refill slots with already-used selection IDs

  NVIDIA/NVFlare#5689 · Filter by author 0z5a opened on Nov 4, 2026

• Add experimental memorization mitigation and audit components for federated fine-tuning

  NVIDIA/NVFlare#5680 · Filter by author holgerroth opened on Nov 3, 2026
  

• Preserve full CA subjects when issuing provisioning certificates

  NVIDIA/NVFlare#5670 · Filter by author pcnudde opened on Nov 2, 2026 · 2.10

• Modernize build & tooling stack

  NVIDIA/NVFlare#5666 · Filter by author Zethson opened on Nov 2, 2026
  7

• Support Python 3.15

  NVIDIA/NVFlare#5664 · Filter by author Zethson opened on Nov 2, 2026

• Research listing: FedMoSA (MICCAI 2026)

  NVIDIA/NVFlare#5652 · Filter by author obedjamir opened on Oct 31, 2026

• [BUG] dict_utils.update raises AttributeError on Python 3.10+ due to collections.Mapping

  NVIDIA/NVFlare#5651 · Filter by author fallintoplace opened on Oct 31, 2026

• Reject job-scoped certificates as job package signers

  NVIDIA/NVFlare#5646 · Filter by author pcnudde opened on Oct 30, 2026 · 2.10

• Add job and study resource statistics

  NVIDIA/NVFlare#5638 · Filter by author nvkevlu opened on Oct 30, 2026

• [BUG] Site config lookups: resources policy ignores resources.conf/.yml, and job folders can shadow site configs

  bug
  NVIDIA/NVFlare#5634 · Filter by author pcnudde opened on Oct 30, 2026
  1

• [BUG] hello-lightning-eval README points to generate_pretrain.py, doesn't exist

  NVIDIA/NVFlare#5624 · Filter by author YujunOh opened on Oct 30, 2026