jellyho/acrft-annot-noprop
acrft-annot-noprop RLT annotation for AC-RFT critic training: raw memmaps (.dat) + meta.json, read directly by scripts/train_rlt_critic.py and scripts/eval_rlt_critic.py in the openpi fork. Shapes are in meta.json: rl_token [T, D], base_action [T, N, H, A], action_chunk [T, H, A], reward/mc_return/done/episode_index/frame_index [T], base_action_heldout [T, num_heldout, H, A]. dtype and reward_scheme are in meta.json. Load with numpy.memmap. from huggingface_hub import… See the full description on the dataset page: https://huggingface.co/datasets/jellyho/acrft-annot-noprop.
acrft-annot-noprop
RLT annotation for AC-RFT critic training: raw memmaps (.dat) + meta.json, read directly by scripts/train_rlt_critic.py and scripts/eval_rlt_critic.py in the openpi fork.
Shapes are in meta.json: rltoken [T, D], baseaction [T, N, H, A], actionchunk [T, H, A], reward/mcreturn/done/episodeindex/frameindex [T], baseactionheldout [T, numheldout, H, A]. dtype and rewardscheme are in meta.json. Load with numpy.memmap.
from huggingfacehub import snapshotdownload d = snapshotdownload("jellyho/acrft-annot-noprop", repotype="dataset") uv run scripts/trainrltcritic.py --data $d --kind arq --steps 100000 ...