SOURCE AND PROVENANCE

Included source snapshot

Source, tests, recipes and documentation for this version. Execution instructions describe local environments; model weights and downloaded data stay on the cluster.

.

.dockerignore.gitignoreMANIFEST.inREADME.mdpyproject.toml

benchmarks

README.md

benchmarks/reasoning_reliability

data.jsonlpredictions-candidate.jsonlpredictions.jsonlspec-candidate.jsonspec.json

benchmarks/tool_workflow_reliability

data.jsonlpredictions-candidate.jsonlpredictions.jsonlspec-candidate.jsonspec.json

cluster/determined

german-qwen-agent-rl-shell.yamlgerman-qwen-openenv-command.yamlgerman-qwen-shell.yamljaif-cpu-command.yamljaif-managed-cpu.yamljaif-managed-openenv.yamljaif-managed-qualification-openenv.yamljaif-managed-qualification.yamljaif-managed-training.yamljaif-shell.yaml

containers

Dockerfile

docs

README.mdagent-gyms-validation.jsonagent-gyms.mdagent-harness.mdarchitecture.mdbenchmark-designs.mddomains.mdenvironment-frameworks.mdevaluation.mdhuggingface-publication.jsonhuggingface-ui-validation.jsonjaif-jobs.mdnemo-gym-stress-validation.jsonopenenv-validation.jsonoperations.mdprogramme-status.mdrelated-work.mdreleases.mdroadmap.mdtraining.mdvalidation.jsonvalidation.md

docs/use-cases

german-agent-rl-results.jsongerman-agent-rl-results.mdgerman-agent-rl.mdgerman-nemo-long-run.mdgerman-openenv-comparison.mdgerman-reasoning-results.jsongerman-reasoning-results.mdgerman-reasoning.md

examples

agent_harness.pyevaluation_cpu_smoke.py

examples/agent_gyms

nemo_gym_client_smoke.pynemo_gym_server.pynemo_gym_smoke.pynemo_gym_stress.pyopenenv_server.pyopenenv_smoke.pyverifiers_environment.pyverifiers_smoke.py

examples/agent_lightning

hessian_lightning_harness.pylaunch.pytrain.py

examples/german-reasoning

DATA.mdreport.py

examples/huggingface

README.mdRUN.md

examples/huggingface/brand

OFL-Poppins.txtPoppins-Bold.woff2Poppins-Light.woff2Poppins-Regular.woff2Poppins-SemiBold.woff2favicon-32x32.pnghessian-ai-logo.svgjaif-logo.pngjaif-provenance.jsonprovenance.jsonsafari-pinned-tab.svg

examples/training

cpu_smoke.pyprompts.jsonlsft.jsonlteacher_sequences.jsonl

recipes

agent-lightning.jsonreasoning-grpo.jsonreasoning-sft.jsonteacher-generation.json

requirements

german-qwen-agent-rl-cu126.txtgerman-qwen-cu126.txtgerman-qwen-managed-cu126.txtinspect.txtnemo-gym.txtopenenv.txttraining-cpu.txtverifiers.txt

scripts

gpu-preflight.pyrun-german-qwen.shrun-nemo-long-queued.pyrun-openenv-after-nemo.pyslurm-train.sbatch

src/hessian_recipes

__init__.py__main__.pyagent_rl_evaluation.pyagent_rl_pilot.pyagent_rl_report.pyagent_rl_tasks.pyagent_rl_training.pybudget_lineage.pycheckpoints.pycli.pydistillation.pyevaluation.pygerman_data.pygerman_pilot.pygerman_report.pygym_client.pygym_experiment.pygym_experiment_evaluation.pygym_training.pygym_training_long.pyhf_recipe_hub.pyhub_content.pyhub_deploy.pyhub_theme.pyinspect_tasks.pymanaged_training.pymanaged_trial.pyopenenv_client.pyopenenv_experiment.pyopenenv_training.pyreleases.pytraining.pytrajectories.py

templates/huggingface

dataset-card.mdmodel-card.mdorganization.md

tests

test_agent_lightning.pytest_agent_rl_evaluation.pytest_agent_rl_pilot.pytest_agent_rl_report.pytest_agent_rl_tasks.pytest_agent_rl_training.pytest_bounded_qwen_parser.pytest_budget_lineage.pytest_checkpoints.pytest_cli.pytest_distillation.pytest_evaluation.pytest_german_data.pytest_german_pilot.pytest_german_report.pytest_gym_client.pytest_gym_experiment.pytest_gym_experiment_evaluation.pytest_gym_recipe_stress.pytest_gym_training.pytest_gym_training_long.pytest_hf_recipe_hub.pytest_hub_deploy.pytest_inspect_tasks.pytest_managed_engine_hooks.pytest_managed_training.pytest_managed_trial.pytest_nemo_queue_launcher.pytest_openenv_client.pytest_openenv_experiment.pytest_openenv_queue.pytest_openenv_server.pytest_openenv_training.pytest_parser_engine_runtime_identity.pytest_releases.pytest_sampled_token_history.pytest_training.pytest_trajectories.pytest_verifiers_gym.py