NousResearch/atropos v0.4.0

v0.4.0

Repository: NousResearch/atropos

Tag: v0.4.0

Published: 2026-03-10T04:20:36Z

Prerelease: no

Release notes:

Highlights

New example trainer

Weights are shared between vLLM and the trainer, no comms needed to sync weights, and memory saved by using only one copy of the weights!

On Policy/Self Distillation Support

Now support logprobs from a teacher/prompted endpoint, fully supporting on policy distillation/self distillation!

OpenAI Endpoint for managed server

Launch an openai endpoint and collect rollouts from any program that takes in an openai endpoint!

What's Changed

• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/215
• Interleaved Tool-Use Within Reasoning Blocks by @interstellarninja in https://github.com/NousResearch/atropos/pull/195
• Pairwise Judgement Environment - improve dataloading, ctx len by @teknium1 in https://github.com/NousResearch/atropos/pull/218
• Add Word Hunt environment by @Aboozle1 in https://github.com/NousResearch/atropos/pull/220
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/222
• qwen tokenizer wrapper & fixed jinja template for tool handling by @shannonsands in https://github.com/NousResearch/atropos/pull/224
• Add arena-hard v1 environment by @teknium1 in https://github.com/NousResearch/atropos/pull/219
• Textworld minimal by @shannonsands in https://github.com/NousResearch/atropos/pull/225
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/228
• Diplomacy trainer env by @shannonsands in https://github.com/NousResearch/atropos/pull/227
• build: update checkout action to v5 by @rejected-l in https://github.com/NousResearch/atropos/pull/233
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/231
• fix: division-by-zero in gradient calculation by @brawncode in https://github.com/NousResearch/atropos/pull/236
• add error logging to collect_trajectories so they don't fail silently by @dmahan93 in https://github.com/NousResearch/atropos/pull/237
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/238
• Update bibtex by @hjc-puro in https://github.com/NousResearch/atropos/pull/235
• Refusalbench v2 by @J-SUPHA in https://github.com/NousResearch/atropos/pull/239
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/241
• Refusalbench v2 by @J-SUPHA in https://github.com/NousResearch/atropos/pull/242
• Fix multiple scored data groups by @shannonsands in https://github.com/NousResearch/atropos/pull/223
• Revert "Fix multiple scored data groups" by @shannonsands in https://github.com/NousResearch/atropos/pull/243
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/246
• fix typo in variable name by @prestoalvarez in https://github.com/NousResearch/atropos/pull/245
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/248
• Multi-Turn Tool-Use RL Environment by @interstellarninja in https://github.com/NousResearch/atropos/pull/160
• WIP: Environments/bleuberi by @aniemerg in https://github.com/NousResearch/atropos/pull/175
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/249
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/251
• refactor(api): improve attribute checking and remove hardcoded values by @DeVikingMark in https://github.com/NousResearch/atropos/pull/250
• fix: correct typos in documentation and comments by @viktorking7 in https://github.com/NousResearch/atropos/pull/254
• [Environment]: smolagents by @aniemerg in https://github.com/NousResearch/atropos/pull/104
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/255
• SmolAgent Env Linting Fixes by @ropresearch in https://github.com/NousResearch/atropos/pull/256
• group temps, sample temps, and logprob api params by @ropresearch in https://github.com/NousResearch/atropos/pull/253
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/257
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/259
• docs: minor fixes to follow code standards by @andrewshab3 in https://github.com/NousResearch/atropos/pull/261
• GZip Compression by @ropresearch in https://github.com/NousResearch/atropos/pull/263
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/266
• docs: few minor fixes by @letmehateu in https://github.com/NousResearch/atropos/pull/265
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/268
• add sglang specific token level logprob handling and server manager/b… by @dmahan93 in https://github.com/NousResearch/atropos/pull/264
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/269
• fix: correct typo and improve code quality by @bobtajson in https://github.com/NousResearch/atropos/pull/267
• add managed vllm server by @dmahan93 in https://github.com/NousResearch/atropos/pull/273
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/275
• refactor: Refactor scored data handling into reusable helper by @ninastef in https://github.com/NousResearch/atropos/pull/272
• feat: dump evaluate subcommand config to YAML in env save dir by @dhyaneesh in https://github.com/NousResearch/atropos/pull/274
• fix some issues by @teknium1 in https://github.com/NousResearch/atropos/pull/279
• docs: fix dead links by @kseniaeremekno in https://github.com/NousResearch/atropos/pull/277
• Convert Environments to ManagedServer for Tinker Integrations by @teknium1 in https://github.com/NousResearch/atropos/pull/278
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/281
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NousResearch/atropos/pull/284
• README updates for Tinker Integration by @samherring99 in…