trace / README.md
maveryn's picture
Add TRACE paper citation
62e343e verified
|
Raw
History Blame Contribute Delete
5.63 kB
metadata
pretty_name: TRACE
language:
  - en
license: cc-by-4.0
annotations_creators:
  - machine-generated
source_datasets:
  - original
task_categories:
  - visual-question-answering
size_categories:
  - 10K<n<100K
tags:
  - visual-reasoning
  - synthetic
  - rlvr
  - multimodal
  - grounded-reasoning
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train/trace_rlvr_train_64000_all1000_seed42-*.parquet
      - split: validation
        path: >-
          data/validation/trace_rlvr_validation_iid_2000_all1000_seed1042.parquet

TRACE: A Taxonomy-Guided Environment for Multidomain Visual Reasoning

TRACE is a synthetic dataset for grounded visual reasoning and reinforcement learning with verifiable rewards. Each example pairs a rendered image with two prompt variants, typed targets, a reward contract, and a reference to its execution trace.

Paper · Project page · GitHub · Collection · Task catalog · Training recipes · 3B model · 7B model

Dataset structure

The release covers 1,000 tasks, 277 scenes, and 11 visual domains: charts, games, geometry, graphs, icons, illustrations, pages, physics, puzzles, symbolic notation, and 3D scenes.

Split Examples Coverage Generation seed
train 64,000 64 examples per task 42
validation 2,000 2 examples per task 1042

Both splits are deterministically shuffled with row-order seed 20260711. Validation is IID over the same 1,000 tasks; it is not a held-out test of new task families.

This is the canonical dataset release for reproducing the public TRACE 3B and 7B checkpoints. The dataset-v1 tag is the stable human-facing release name; the public training configurations pin its immutable commit. An independent equivalence receipt verifies that every field consumed by training is identical to the original run input, including embedded image bytes and row order.

The optional trace_supervision_mode column is advisory metadata. The released training profiles select prompt_answer, score answer_gt, and do not read this column.

Fields

Field Description
images, image_sizes Embedded rendered image and dimensions defining the annotation coordinate frame
prompt_answer Prompt selected by the released training profiles
prompt_answer_and_annotation Prompt requesting an answer and image-space annotation
answer_gt, annotation_gt JSON-encoded typed verifier targets
reward_contract JSON-encoded metadata-backed scoring contract
instance_id, domain, scene_id, task, query_id Identity and public taxonomy
trace_ref JSON-encoded reference to the compressed execution-trace sidecar
trace_supervision_mode Optional advisory prompt-selection recommendation; unused by the released training profiles

The sidecars/ directories contain build and validation reports, canonical instance records, curriculum indices, manifests, and compressed execution traces. Images are embedded in the Parquet files and are not duplicated in the sidecars. Report and manifest paths on the default branch are repository-relative; dataset ids and content hashes retain artifact identity.

Loading

from datasets import load_dataset

dataset = load_dataset("maveryn/trace", revision="dataset-v1")
train = dataset["train"]
validation = dataset["validation"]

The contract fields are stored as JSON strings. Decode them before scoring:

import json

row = dataset["train"][0]
answer = json.loads(row["answer_gt"])
annotation = json.loads(row["annotation_gt"])
reward_contract = json.loads(row["reward_contract"])

The public 3B and 7B training configurations pin the immutable release commit and validate the split sizes, task coverage, schema, prompts, and model inputs before training.

Intended use and limitations

TRACE is intended for research on multimodal reasoning, verifiable-reward post-training, grounded supervision, synthetic curricula, and task-level analysis. Its procedurally generated images do not represent the full distribution of natural images or real documents. Prompts are predominantly English, and metadata-backed labels do not eliminate model or task-design biases. The dataset should not be treated as evidence of real-world model reliability or safety.

Citation

If you use TRACE, please cite:

@misc{alam2026trace,
  title         = {Trace: A Taxonomy-Guided Environment for Multidomain Visual Reasoning},
  author        = {Alam, Md Tanvirul},
  year          = {2026},
  eprint        = {2607.19790},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  url           = {https://arxiv.org/abs/2607.19790}
}

License

The dataset is released under CC BY 4.0. The generator software is separately licensed under Apache-2.0; bundled assets and dependencies retain the terms documented in the source repository.