8000 Advance development after v0.12.0 release by DaoyuanLi2816 · Pull Request #114 · DaoyuanLi2816/mini-verl · GitHub
[go: up one dir, main page]

Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion CITATION.cff
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@ title: "miniVERL: A bounded single-GPU runtime for verl-style OPD"
message: "If you use miniVERL in your work, please cite it as below."
type: software
version: 0.12.0
date-released: 2026-09-07
date-released: 2026-09-08
license: Apache-2.0
repository-code: "https://github.com/DaoyuanLi2816/mini-verl"
url: "https://github.com/DaoyuanLi2816/mini-verl"
Expand Down
10 changes: 5 additions & 5 deletions PROJECT_STATE.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,21 +4,21 @@ Current maintainer handoff for **miniVERL** (`mini-verl` package, `miniverl`
CLI). `release-state.yaml` is the canonical version source; this page indexes
current product and evidence state rather than repeating release history.

Last updated: 2026-09-07.
Last updated: 2026-09-08.

Canonical release state: releasing `v0.12.0`.
Canonical release state: stable `v0.12.0` (`94216828ecab77d5f87a541563f5c2d8951d5a33`), development `0.12.1.dev0`.

## Release state

- Release candidate: `v0.12.0`; publication is gated on exact-SHA RTX 4080 qualification.
- Previous stable release: `v0.11.0` at `479695c4980150bc6f7a51bddec2b90c08c8bb98`.
- Stable release: `v0.12.0` at `94216828ecab77d5f87a541563f5c2d8951d5a33`.
- Development line: `0.12.1.dev0`.
- Stable docs: <https://daoyuanli2816.github.io/mini-verl/>.
- Development docs: <https://daoyuanli2816.github.io/mini-verl/dev/>.
- Historical build log: [v0.1-v0.9 archive](docs/history/project-state-v0.1-v0.9.md).

## Current product boundary

Release `0.12.0` has a closed typed profile compiler and adds
Development `0.12.1.dev0` has a closed typed profile compiler and adds
`verl-rl-v0.9-single-gpu-v1`, a typed resolved
subset of official verl `v0.9.0` at commit
`483b8a009ba3a97563edee3a19887e4862b8094a`. It compiles GRPO, Dr.GRPO, RLOO
Expand Down
46 changes: 23 additions & 23 deletions PYPI.md
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
<p align="center">
<img src="https://raw.githubusercontent.com/DaoyuanLi2816/mini-verl/v0.12.0/docs/banner.svg" alt="miniVERL — lower verl experiment semantics onto one CUDA GPU" width="880">
<img src="https://raw.githubusercontent.com/DaoyuanLi2816/mini-verl/main/docs/banner.svg" alt="miniVERL — lower verl experiment semantics onto one CUDA GPU" width="880">
</p>

<div align="center">
Expand All @@ -8,15 +8,15 @@
[![Build](https://github.com/DaoyuanLi2816/mini-verl/actions/workflows/build.yml/badge.svg)](https://github.com/DaoyuanLi2816/mini-verl/actions/workflows/build.yml)
[![PyPI](https://img.shields.io/pypi/v/miniverl.svg)](https://pypi.org/project/miniverl/)
[![Python](https://img.shields.io/badge/python-3.10%20%7C%203.11%20%7C%203.12%20%7C%203.13-blue)](https://www.python.org)
[![License](https://img.shields.io/badge/license-Apache--2.0-blue)](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/LICENSE)
[![License](https://img.shields.io/badge/license-Apache--2.0-blue)](https://github.com/DaoyuanLi2816/mini-verl/blob/main/LICENSE)

</div>

<p align="center">
<a href="https://pypi.org/project/miniverl/"><strong>PyPI</strong></a> ·
<a href="https://daoyuanli2816.github.io/mini-verl/"><strong>Stable docs</strong></a> ·
<a href="https://daoyuanli2816.github.io/mini-verl/dev/">Development docs</a> ·
<a href="https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/README.zh-CN.md">中文</a>
<a href="https://github.com/DaoyuanLi2816/mini-verl/blob/main/README.zh-CN.md">中文</a>
</p>

**miniVERL runs a validated subset of verl experiment semantics on one NVIDIA
Expand Down Expand Up @@ -51,7 +51,7 @@ load the pinned Qwen3-0.6B actor and execute the two-iteration example.

The `[train,cuda]` extra installs the ML and quantization stack; select PyTorch's
CUDA build separately with the [PyTorch installer](https://pytorch.org/get-started/locally/).
The [single-GPU guide](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/single-gpu-guide.md) covers memory planning and the
The [single-GPU guide](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/single-gpu-guide.md) covers memory planning and the
maintainer-measured RTX 4080 environment.

## What a run gives you
Expand All @@ -70,8 +70,8 @@ maintainer-measured RTX 4080 environment.
## How it works

<picture>
<source media="(max-width: 640px)" srcset="https://raw.githubusercontent.com/DaoyuanLi2816/mini-verl/v0.12.0/docs/verl-local-runtime-mobile.svg">
<img src="https://raw.githubusercontent.com/DaoyuanLi2816/mini-verl/v0.12.0/docs/verl-local-runtime.svg" alt="A resolved verl config compiles into a validated single-GPU execution plan; actor, reference, teacher and reward roles execute in phases and produce portable artifacts plus a readiness report.">
<source media="(max-width: 640px)" srcset="https://raw.githubusercontent.com/DaoyuanLi2816/mini-verl/main/docs/verl-local-runtime-mobile.svg">
<img src="https://raw.githubusercontent.com/DaoyuanLi2816/mini-verl/main/docs/verl-local-runtime.svg" alt="A resolved verl config compiles into a validated single-GPU execution plan; actor, reference, teacher and reward roles execute in phases and produce portable artifacts plus a readiness report.">
</picture>

One GPU is treated as a temporal scheduler for logical roles. Critic-free RL
Expand All @@ -84,10 +84,10 @@ with a teacher-scoring phase in place of outcome-only rewards.

| Goal | First command | Primary artifact | Next step |
| --- | --- | --- | --- |
| **Run local RL** | `miniverl import-verl --profile verl-rl-v0.9-single-gpu-v1 --config verl-rl.yaml --out local.yaml` | native recipe + compatibility report | [RL quickstart](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/verl-rl-runtime.md) |
| **Run local OPD** | `miniverl plan --profile verl-opd-v0.8-single-gpu-v1 --config verl-opd.yaml --out plan.json` | immutable execution plan | [OPD quickstart](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/opd-quickstart.md) |
| **Fit your GPU** | `miniverl plan --config verl-opd.yaml --probe` | measured placement plan | [Hardware planning](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/hardware-planning.md) |
| **Hand off artifacts** | `miniverl export-verl --run runs/my-run --target-verl v0.8.0 --out scaleout` | PEFT + Parquet + config bundle | [Scale-out contract](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/verl-opd-scaleout.md) |
| **Run local RL** | `miniverl import-verl --profile verl-rl-v0.9-single-gpu-v1 --config verl-rl.yaml --out local.yaml` | native recipe + compatibility report | [RL quickstart](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/verl-rl-runtime.md) |
| **Run local OPD** | `miniverl plan --profile verl-opd-v0.8-single-gpu-v1 --config verl-opd.yaml --out plan.json` | immutable execution plan | [OPD quickstart](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/opd-quickstart.md) |
| **Fit your GPU** | `miniverl plan --config verl-opd.yaml --probe` | measured placement plan | [Hardware planning](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/hardware-planning.md) |
| **Hand off artifacts** | `miniverl export-verl --run runs/my-run --target-verl v0.8.0 --out scaleout` | PEFT + Parquet + config bundle | [Scale-out contract](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/verl-opd-scaleout.md) |

Native recipes also support SFT, DPO and offline KD, plus calculator, JSON
navigation, read-only SQLite and custom tool environments.
Expand All @@ -105,7 +105,7 @@ navigation, read-only SQLite and custom tool environments.
| PPO / GAE execution | not implemented | math is conformance-tested; a trainable critic lifecycle is not yet present |
| Ray, FSDP/FSDP2, Megatron, TP/PP/DP > 1 | distributed-only | these change physical scale, not the local objective |

The generated [v0.9 compatibility record](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/generated/verl-rl-v0.9-compatibility.json)
The generated [v0.9 compatibility record](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/generated/verl-rl-v0.9-compatibility.json)
binds every example field to its source value, local target, classification and
compiler-rule digest. `miniverl import-verl` accepts a resolved documented
subset, rejects unknown algorithm-changing fields, and never substitutes a
Expand All @@ -117,31 +117,31 @@ The published Qwen3-0.6B/1.7B OPD workload consumed 32 distinct prompts and
completed eight current-policy updates at 3.1914 GiB peak reserved VRAM on an
RTX 4080. A matched interruption after update four reproduced byte-identical
trajectories, adapter and optimizer tensors. The SmolLM2-360M/1.7B workload
completed the same shape at 1.4961 GiB. [System records](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/verl-opd-reference-workload.md)
completed the same shape at 1.4961 GiB. [System records](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/verl-opd-reference-workload.md)
contain configs, hashes and phase timings.

Rollout Runtime v2 measured `hf_cached` and managed vLLM across 24 RTX 4080
cells spanning response length, sampling mode and `n=1/4`. See the
[runtime report](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/benchmarks/rollout-runtime-v2.md). Evidence for the new
[runtime report](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/benchmarks/rollout-runtime-v2.md). Evidence for the new
RL family is published through the exact-wheel release qualification rather
than presented as a task-quality comparison.

## Research record

The scientific reports preserve their original outcomes and frozen inputs:
[calculator protocol](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/benchmarking.md),
[RecoveryBench](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/recoverybench/recoverybench-v1.md),
[Alignment Lab](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/alignment-lab/alignment-lab-v1.md), and the
[External Alignment Gate](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/alignment-external/alignment-external-v1.md).
[calculator protocol](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/benchmarking.md),
[RecoveryBench](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/recoverybench/recoverybench-v1.md),
[Alignment Lab](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/alignment-lab/alignment-lab-v1.md), and the
[External Alignment Gate](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/alignment-external/alignment-external-v1.md).
They are scoped studies, separate from runtime qualification.

## Compatibility boundary

miniVERL targets one local process and one NVIDIA CUDA GPU. The compiler
distinguishes exact or conformant semantics, local physical lowering,
distributed-only settings, feasible work not yet implemented, and technically
unsupported inputs. The [compatibility policy](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/compatibility.md) is the
complete matrix; [limitations](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/limitations.md) collects architecture,
unsupported inputs. The [compatibility policy](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/compatibility.md) is the
complete matrix; [limitations](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/limitations.md) collects architecture,
measurement, security and generalization boundaries in one place. miniVERL is
an independent Apache-2.0 project.

Expand All @@ -154,7 +154,7 @@ python -m pip install -e ".[dev,train]"
pytest -q -m "not gpu and not network"
```

See [CONTRIBUTING.md](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/CONTRIBUTING.md), [CHANGELOG.md](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/CHANGELOG.md),
[CITATION.cff](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/CITATION.cff), the [guide for verl users](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/docs/for-verl-users.md),
[SECURITY.md](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/SECURITY.md) and the
[Apache-2.0 license](https://github.com/DaoyuanLi2816/mini-verl/blob/v0.12.0/LICENSE).
See [CONTRIBUTING.md](https://github.com/DaoyuanLi2816/mini-verl/blob/main/CONTRIBUTING.md), [CHANGELOG.md](https://github.com/DaoyuanLi2816/mini-verl/blob/main/CHANGELOG.md),
[CITATION.cff](https://github.com/DaoyuanLi2816/mini-verl/blob/main/CITATION.cff), the [guide for verl users](https://github.com/DaoyuanLi2816/mini-verl/blob/main/docs/for-verl-users.md),
[SECURITY.md](https://github.com/DaoyuanLi2816/mini-verl/blob/main/SECURITY.md) and the
[Apache-2.0 license](https://github.com/DaoyuanLi2816/mini-verl/blob/main/LICENSE).
34 changes: 17 additions & 17 deletions docs/generated/quality.json
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
{
"schema_version": 2,
"release": "0.12.0",
"status": "candidate",
"status": "released",
"quality_floor": "2,000+ tests and 80%+ branch coverage at v0.12.0",
"local_validation": {
"scope": "the maintainer's workstation, including Windows CUDA and network paths plus the exact official verl v0.9 checkout",
Expand All @@ -28,25 +28,25 @@
},
"release_validation": {
"scope": "the exact v0.12.0 release commit",
"commit": "pending",
"commit": "94216828ecab77d5f87a541563f5c2d8951d5a33",
"workflows": {
"ci": "pending",
"build": "pending",
"docs": "pending",
"pinned_verl_bridge": "pending",
"gpu_full_qualification": "pending",
"release_dry_run": "pending",
"release": "pending"
"ci": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34193726077",
"build": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34193726069",
"docs": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34198700305",
"pinned_verl_bridge": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34198700339",
"gpu_full_qualification": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34193785555",
"release_dry_run": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34197998778",
"release": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34198700279"
},
"conclusion": "pending",
"gpu_coverage": "pending exact-SHA attempt-1 full qualification on the WSL2 RTX 4080 runner",
"conclusion": "passed",
"gpu_coverage": "attempt-1 full qualification of the exact release wheel on one WSL2 RTX 4080; 16 GRPO trajectories and two finite parameter-changing updates passed alongside the v0.11 regression matrix",
"publication": {
"pypi": "pending",
"github_release": "pending",
"documentation": "pending",
"immutable_documentation_build": "pending",
"wheel_sha256": "pending",
"sdist_sha256": "pending"
"pypi": "https://pypi.org/project/miniverl/0.12.0/",
"github_release": "https://github.com/DaoyuanLi2816/mini-verl/releases/tag/v0.12.0",
"documentation": "https://daoyuanli2816.github.io/mini-verl/",
"immutable_documentation_build": "https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34198700305",
"wheel_sha256": "41a7d59b792b977bc4d81d644fc4429e303493b3c4417a6c213c68ad7b0b345f",
"sdist_sha256": "38b6f77f27e3986867b5d453c684f7a7b15205809643cfc8470a45e02e3e84ff"
}
}
}
4 changes: 2 additions & 2 deletions docs/overrides/main.html
Original file line number Diff line number Diff line change
@@ -1,12 +1,12 @@
{% extends "base.html" %}

{% block announce %}
<div class="docs-channel" data-stable-version="0.12.0" data-dev-version="0.12.0">
<div class="docs-channel" data-stable-version="0.12.0" data-dev-version="0.12.1.dev0">
<strong id="docs-channel-label">Stable documentation</strong>
<label for="docs-version-selector">Version</label>
<select id="docs-version-selector" aria-label="Documentation version">
<option value="stable">Stable 0.12.0</option>
<option value="dev">Development 0.12.0</option>
<option value="dev">Development 0.12.1.dev0</option>
</select>
</div>
{% endblock %}
28 changes: 24 additions & 4 deletions docs/release-checklist.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,11 @@ This is the release gate and publication record for miniVERL. A checked item
names an invariant exercised on the stated source. Publication begins only
after the exact release commit and its remote checks are green.

## v0.12.1 development

- [x] Begin from the verified v0.12.0 release and advance only the canonical
development state to `0.12.1.dev0`.

## v0.12.0 development

- [x] Pin the critic-free RL compiler and conformance fixtures to upstream verl
Expand Down Expand Up @@ -60,11 +65,26 @@ after the exact release commit and its remote checks are green.

## Publication record

### v0.12.0 pending publication
### v0.12.0

- [ ] Record the exact-release-SHA GPU qualification, dry run and tag workflow.
- [ ] Verify PyPI, GitHub Release, stable/dev docs and a clean public install.
- [ ] Advance main to `0.12.1.dev0` in a state-sync PR.
- [x] Exact-release-SHA WSL2 full qualification run
[`34193785555`](https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34193785555)
passed on attempt 1 for commit
`94216828ecab77d5f87a541563f5c2d8951d5a33`.
- [x] Non-publishing release dry run
[`34197998778`](https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34197998778)
accepted the same candidate and full qualification without rebuilding.
- [x] Tag workflow
[`34198700279`](https://github.com/DaoyuanLi2816/mini-verl/actions/runs/34198700279)
verified PyPI hashes, trusted-publisher attestations and a clean public
install before creating the
[v0.12.0 GitHub Release](https://github.com/DaoyuanLi2816/mini-verl/releases/tag/v0.12.0).
- [x] PyPI and the GitHub Release expose the identical wheel at SHA-256
`41a7d59b792b977bc4d81d644fc4429e303493b3c4417a6c213c68ad7b0b345f`
and sdist at SHA-256
`38b6f77f27e3986867b5d453c684f7a7b15205809643cfc8470a45e02e3e84ff`.
- [x] This state-sync PR advances main to `0.12.1.dev0`; merge only after its
required checks pass.

- [x] Exact-release-SHA WSL2 full qualification run
[`33288192471`](https://github.com/DaoyuanLi2816/mini-verl/actions/runs/33288192471)
Expand Down
8 changes: 4 additions & 4 deletions release-state.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -17,13 +17,13 @@
# such distinction, which is how its tag shipped a docs selector still
# advertising "Stable 0.6.1 / Development 0.6.2.dev0".
schema_version: 1
phase: release
phase: development

stable:
version: "0.12.0"
tag: "v0.12.0"
release_commit: pending
released_at: "2026-09-07"
release_commit: "94216828ecab77d5f87a541563f5c2d8951d5a33"
released_at: "2026-09-08"

development:
version: "0.12.0"
version: "0.12.1.dev0"
2 changes: 1 addition & 1 deletion src/miniverl/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,6 @@

from __future__ import annotations

__version__ = "0.12.0"
__version__ = "0.12.1.dev0"

__all__ = ["__version__"]
0