Datasetpublic

ParoleLM/socialjax-harvest-frame-aligned-512

SocialJax Harvest frame-aligned dynamics dataset Compact tokenizer-code dataset for training action-conditioned SocialJax Harvest dynamics models. Repository: ParoleLM/socialjax-harvest-frame-aligned-512 Format: frame_aligned_socialjax_dynamics_v3 Tokenizer codes per frame: 88 Codebook size: 512 Maximum agents: 7 Total size: 4.385 GB Train: 9,720 rollouts, 19,916,280 frames Validation: 540 rollouts, 1,106,460 frames Test: 540 rollouts, 1,106,460 frames Files… See the full description on the dataset page: https://huggingface.co/datasets/ParoleLM/socialjax-harvest-frame-aligned-512.

sourceHugging Faceupdated 4mo agoView on Hugging Face

0likes16downloads

Dataset Card

SocialJax Harvest frame-aligned dynamics dataset

Compact tokenizer-code dataset for training action-conditioned SocialJax Harvest dynamics models.

—Repository: ParoleLM/socialjax-harvest-frame-aligned-512
—Format: frame_aligned_socialjax_dynamics_v3
—Tokenizer codes per frame: 88
—Codebook size: 512
—Maximum agents: 7
—Total size: 4.385 GB
—Train: 9,720 rollouts, 19,916,280 frames
—Validation: 540 rollouts, 1,106,460 frames
—Test: 540 rollouts, 1,106,460 frames

Files

Each split contains:

—tokens.bin: uint16[num_frames,88]
—actions.bin: int8[num_transitions,7]
—rewards.bin: float16[num_transitions,7]
—dones.bin: bool[num_transitions]
—docs.jsonl: rollout boundaries, frame/transition offsets, active-agent mask, and provenance

Action a_t causes transition z_t -> z_{t+1}. Training windows must stay within one rollout document. There is no EOS token. See meta.json for exact split sizes and tokenizer provenance.