Releases: DaoyuanLi2816/mini-verl
Releases · DaoyuanLi2816/mini-verl
Release list
miniVERL v0.10.0
What's Changed
- Advance development to 0.10.0.dev0 by @DaoyuanLi2816 in #78
- Add versioned compatibility profile registry by @DaoyuanLi2816 in #79
- Add measured sampled-k1 OPD profile by @DaoyuanLi2816 in #80
- Measure the full SmolLM2 OPD recipe by @DaoyuanLi2816 in #81
- Add portable hardware evidence records by @DaoyuanLi2816 in #82
- Release miniVERL v0.10.0 by @DaoyuanLi2816 in #83
Full Changelog: v0.9.1...v0.10.0
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.9.1
What's Changed
- Advance development to 0.9.1.dev0 by @DaoyuanLi2816 in #74
- Repair verl OPD semantic contracts by @DaoyuanLi2816 in #75
- Align verl OPD documentation with runtime truth by @DaoyuanLi2816 in #76
- Release v0.9.1 by @DaoyuanLi2816 in #77
Full Changelog: v0.9.0...v0.9.1
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.9.0
What's Changed
- Advance development to 0.9.0.dev0 by @DaoyuanLi2816 in #67
- Add verl-style override workflow by @DaoyuanLi2816 in #68
- Bind runs to immutable execution plans by @DaoyuanLi2816 in #69
- Add bounded CUDA hardware probe by @DaoyuanLi2816 in #70
- Materialize launchable verl OPD bundles by @DaoyuanLi2816 in #71
- Publish the realistic RTX 4080 OPD workload by @DaoyuanLi2816 in #72
- Prepare the v0.9.0 release by @DaoyuanLi2816 in #73
Full Changelog: v0.8.1...v0.9.0
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.8.1
What's Changed
- Advance development to 0.8.1.dev0 by @DaoyuanLi2816 in #63
- Make the verl OPD workflow product-first by @DaoyuanLi2816 in #65
- Release miniVERL v0.8.1 by @DaoyuanLi2816 in #66
Full Changelog: v0.8.0...v0.8.1
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.8.0
What's Changed
- Advance development to 0.8.0.dev0 by @DaoyuanLi2816 in #57
- Add typed verl v0.8 OPD config compiler by @DaoyuanLi2816 in #58
- Add first-class Parquet prompt OPD runtime by @DaoyuanLi2816 in #59
- Conform forward top-k loss with verl v0.8 by @DaoyuanLi2816 in #60
- Add the single-GPU verl-style OPD plan and run path by @DaoyuanLi2816 in #61
- Release v0.8.0 single-GPU verl OPD runtime by @DaoyuanLi2816 in #62
Full Changelog: v0.7.1...v0.8.0
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.7.1
What's Changed
- Start v0.7.1 development and record v0.7.0 release by @DaoyuanLi2816 in #54
- Correct published evidence digest labels by @DaoyuanLi2816 in #55
- Release v0.7.1 product correction by @DaoyuanLi2816 in #56
Full Changelog: v0.7.0...v0.7.1
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.7.0
What's Changed
- Start v0.6.4 development and record the v0.6.3 release evidence by @DaoyuanLi2816 in #47
- Strip a UTF-8 byte-order mark before parsing a reward scaffold by @DaoyuanLi2816 in #48
- v0.7.0 foundation: refuse redirected bundles and bind sidecars to their dataset by @DaoyuanLi2816 in #49
- v0.7.0 external benchmarks: pinned endpoints, judges and the frozen evaluation suite by @DaoyuanLi2816 in #50
- v0.7.0 preregistration: freeze the study and record a preregistered checkpoint-selection failure by @DaoyuanLi2816 in #51
- Record the preregistration merge commit in the handoff by @DaoyuanLi2816 in #52
- Release v0.7.0 early-stop evidence by @DaoyuanLi2816 in #53
Full Changelog: v0.6.3...v0.7.0
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.6.3
What's Changed
- Start v0.6.3 development and record the v0.6.2 release evidence by @DaoyuanLi2816 in #44
- v0.6.3 security, artifact integrity and release-state hardening by @DaoyuanLi2816 in #45
- Close the v0.6.3 reward-verifier, trust-boundary, sidecar and streaming gaps by @DaoyuanLi2816 in #46
Full Changelog: v0.6.2...v0.6.3
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.6.2
What's Changed
- Advance development to 0.6.2.dev0 by @DaoyuanLi2816 in #41
- Serialize the Pages deploy job across refs by @DaoyuanLi2816 in #42
Full Changelog: v0.6.1...v0.6.2
Contributors
DaoyuanLi2816
Assets 5
miniVERL v0.6.1
What's Changed
- Start v0.6.1 development by @DaoyuanLi2816 in #38
- v0.6.1 visual integrity and bridge correctness by @DaoyuanLi2816 in #40
Full Changelog: v0.6.0...v0.6.1
Contributors
DaoyuanLi2816