2.7.2rc10: Bug fixes and feature improvements
Pre-release
Pre-release
What's Changed
- [2.7] Fix hello-numpy-cross-val example by @nvkevlu in #4168
- [2.7] Add end-to-end download starvation test for stream pool fix by @chesterxgchen in #4172
- [2.7] Fix numpy cross val path of results by @nvkevlu in #4173
- [2.7] update class_path for llm example by @ZiyueXu77 in #4175
- [2.7] Remove roadmap from documentation [skip ci] by @chesterxgchen in #4179
- [2.7] Initial checkpoint info by @holgerroth in #4178
- [2.7] Fix cifar10 integration tests by @YuanTingHsieh in #4150
- [2.7] Avoid self-message deadlock for local swarm result submission by @pcnudde in #4186
- [2.7] smaller lock in produce item by @pcnudde in #4174
- [2.7] Replace job.to with alternative recipe apis by @ZiyueXu77 in #4183
- [2.7] Fix recipes and job api model/initial_model confusion by @YuanTingHsieh in #4188
- [2.7] Fix numpy cross val sticky property by @nvkevlu in #4181
- [2.7] Use Initial Global Model in BioNeMo Recipe Examples by @holgerroth in #4189
- [2.7] Add PR-4172 style tests for swarm self-result submission fix by @pcnudde in #4190
- [2.7] update to match fix in main by @nvkevlu in #4202
- [2.7] Fix RxTask self-deadlock on stream error cleanup by @pcnudde in #4204
- [2.7] Mitigate F3 streaming Head-of-Line (HOL) stalls and add guardrails by @chesterxgchen in #4206
- [2.7] Client-side memory management by @chesterxgchen in #4211
- [2.7] Pass-Through: Zero Tensor Copy at CJ for Large-Model Federated Training by @chesterxgchen in #4210
Full Changelog: 2.7.2rc9...2.7.2rc10