| --- |
| dataset_info: |
| features: |
| - name: messages |
| list: |
| - name: content |
| dtype: string |
| - name: role |
| dtype: string |
| - name: ground_truth |
| dtype: string |
| - name: dataset |
| dtype: string |
| - name: env_config |
| struct: |
| - name: env_name |
| dtype: string |
| - name: image |
| dtype: string |
| - name: task_id |
| dtype: string |
| - name: source |
| dtype: string |
| splits: |
| - name: train |
| num_bytes: 133363912 |
| num_examples: 37484 |
| download_size: 29005927 |
| dataset_size: 133363912 |
| configs: |
| - config_name: default |
| data_files: |
| - split: train |
| path: data/train-* |
| license: cc-by-4.0 |
| language: |
| - en |
| tags: |
| - tmax |
| - terminal |
| - agents |
| - reinforcement-learning |
| --- |
| |
| # Recursive-Task-Synthesis for tmax |
|
|
| **Build images locally:** the complete dataset and build contexts are retained. The hosted SIF image cache has been retired; see the rebuilding instructions below. |
|
|
| All **37,484** task directories from [Zhongzhi1228/Recursive-Task-Synthesis](https://huggingface.co/datasets/Zhongzhi1228/Recursive-Task-Synthesis), pinned to `be44f96808d5a9b599d5cb024341ff00091adeb7`, converted to tmax's `swerl_vanillux_sandbox` format. |
|
|
| The train split uses the same `messages`, `ground_truth`, `dataset`, `env_config`, and `source` schema as the other `hamishivi/agent-task-*` datasets. Messages use the tmax Vanillux templates; `dataset` is `passthrough`. Task IDs are prefixed with `recursive_task_synthesis__` to avoid collisions when combining task archives. `task-manifest.json` maps them back to original paths. |
|
|
| `task-data.tar.gz` contains a directory per task with all original task files preserved byte-for-byte, plus `image.txt`. Instructions, verifiers, build contexts, metadata, and reference solutions are retained. The tmax loader exposes only environment seed files and defers tests until submission; reference solutions are not mounted into the agent sandbox. |
|
|
| ## Usage |
|
|
| ```python |
| from datasets import load_dataset |
| from huggingface_hub import hf_hub_download |
| |
| train = load_dataset("hamishivi/agent-task-recursive-task-synthesis", split="train") |
| archive = hf_hub_download("hamishivi/agent-task-recursive-task-synthesis", "task-data.tar.gz", repo_type="dataset") |
| ``` |
|
|
| Extract the archive and set the sandbox `task_data_dir` to the directory containing the task directories. Every row specifies its image explicitly. Task resource requirements and verifier timeouts remain in the original `task.toml`; configure the training harness accordingly. |
|
|
| ## Images |
|
|
| Images use `hamishi740/agent-task-recursive-task-synthesis:<source-environment-hash>` and target Linux AMD64. See `task-manifest.json` for build status. |
|
|
| ## Validation and attribution |
|
|
| All rows passed schema and Parquet round-trip checks, and every archived task file was checked against the downloaded source bytes. See `validation.json`. This does not constitute an evaluation of every task or a full training run. |
|
|
| The upstream license is cc-by-4.0. Consult the upstream dataset card for citation and provenance. tmax Vanillux prompts are adapted from mini-swe-agent (MIT). |
|
|
| ## Reproducibility and runtime coverage |
|
|
| This conversion targets `hamishivi/tmax-private`, branch `geomean_mask`, commit `9fff7af6e86f2589830637772ff5b10b1c080131`. All original task files are preserved; added task IDs, training rows, and image references adapt the packaging for tmax. See `UPSTREAM_README.md` for the source authors' attribution and citation details. |
|
|
| Validation covers all task records and archived file contents. No container build or runtime task evaluation was performed for this release. Upstream validation claims do not replace testing in your training harness. |
|
|
| ## Resumable image builds |
|
|
| Requires Docker Buildx and a Docker Hub login with write access to the target image repository. The script skips published tags and digest-pinned upstream images, saves per-image logs, and retries unfinished builds on the next run. |
|
|
| ```sh |
| mkdir -p tasks |
| tar -xzf task-data.tar.gz -C tasks |
| python build_dataset_images.py --task-data-dir tasks --manifest task-manifest.json --workers 8 --timeout 1800 |
| ``` |
|
|
| Images built by this script target Linux AMD64. If using another Docker Hub namespace, update image references consistently in the manifest, training records, and task `image.txt` files. The image status file is a publication-time snapshot. |
|
|
| The source includes 67 references to three unavailable local image names. These tasks use the supplied build contexts and new image targets; the original names are retained in the manifest as `unavailable_upstream_image`. Source task-group IDs are also retained for grouping and splitting related tasks. |
|
|
|
|
| ## Hosted image pool retired |
|
|
| The hosted Apptainer SIF cache was removed on **2026-10-19** to reduce repository storage. All **37,484 task records**, original task files, verifiers, build contexts, and attribution remain available in this dataset. Prebuilt SIF downloads and the old image download revisions are no longer available. |
|
|
| The original `apptainer/manifest.jsonl` files are preserved byte-for-byte as historical build records. The complete 29,501-image build provenance, including Docker runtime configuration and recorded repairs, is in [the TMaxxx manifest](https://huggingface.co/datasets/TMaxxx/agent-task-recursive-task-synthesis/blob/main/apptainer/manifest.jsonl). Their `status`, `hf_path`, and `commit` fields describe the retired publication and are no longer active download locations. `apptainer/status.json` marks the pool as retired and every environment as requiring a local build. |
|
|
| ### Rebuild images locally |
|
|
| Download `task-data.tar.gz`, `task-manifest.json`, and `build_dataset_images.py` from this dataset, then run: |
|
|
| ```sh |
| mkdir -p tasks |
| tar -xzf task-data.tar.gz -C tasks |
| python build_dataset_images.py --task-data-dir tasks --manifest task-manifest.json \ |
| --load --max-images 1 --log-dir local-image-build-logs |
| ``` |
|
|
| `--load` builds Linux AMD64 images into your local Docker engine without a Docker Hub login. Start with one image; omit `--max-images 1` when ready to build all environments. The existing push behavior remains available by omitting `--load` and using a registry account with permission to push the requested image names. |
|
|
| The original Dockerfiles may need the recorded build compatibility repairs in `context_repair`, such as archived Debian packages or CentOS mirrors. Rebuilding does not guarantee the exact bytes of the retired SIFs. Task fixtures and verifiers remain unchanged. |
|
|
| To use a locally built Docker image with Apptainer, use `image` and `sif_name` from the environment manifest: |
|
|
| ```sh |
| mkdir -p sifs |
| docker save "$IMAGE" -o image.tar |
| apptainer build "./sifs/$SIF_NAME" "docker-archive://$PWD/image.tar" |
| export SWERL_APPTAINER_SIF_DIR="$PWD/sifs" |
| ``` |
|
|
