The dataset viewer is not available for this split.
Error code: FeaturesError
Exception: FileNotFoundError
Message: [Errno 2] No such file or directory: '<datasets.utils.file_utils.FilesIterable object at 0x7f4b513d7410>'
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/split/first_rows.py", line 247, in compute_first_rows_from_streaming_response
iterable_dataset = iterable_dataset._resolve_features()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 4196, in _resolve_features
features = _infer_features_from_batch(self.with_format(None)._head())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2533, in _head
return next(iter(self.iter(batch_size=n)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2711, in iter
for key, pa_table in ex_iterable.iter_arrow():
^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2249, in _iter_arrow
yield from self.ex_iterable._iter_arrow()
File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 494, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 384, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/xml/xml.py", line 67, in _generate_tables
with open(file, encoding=self.config.encoding, errors=self.config.encoding_errors) as f:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/streaming.py", line 73, in wrapper
return function(*args, download_config=download_config, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/site-packages/datasets/utils/file_utils.py", line 967, in xopen
return open(main_hop, mode, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: '<datasets.utils.file_utils.FilesIterable object at 0x7f4b513d7410>'Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Manipulation VLA 评测全集 (GR00T N1.7 · pi0 · Genie-Envisioner)
ManiSkill 仿真上对多个 VLA 模型做的 pairwise-OOD / all-factor / conflict / VLM 评测的 可复现全集:代码 + 仿真 + 全部评测结果(逐 ckpt 逐 seed)+ 复现步骤。不含 ckpt 权重(见下方 HF 链接)。
怎么跑 → 见 REPRODUCE.md。
目录结构
README.md ← 本文件:总览 + 结果 + 导航
REPRODUCE.md ← 怎么跑每一类评测(环境/ckpt/命令)
code/ ← 全部评测代码(client / runner / dispatcher)
results/
├─ gr00t/
│ ├─ pairgrid_hard/ 6 实验 × 9 ckpt,硬口径(同色同形干扰项)
│ │ {color_size, color_spatial, verb_spatial, spatial_size, spatial_object, verb_size}/
│ ├─ pairgrid_easy/color_size/ 软口径对照(1 干扰项, ~2× 尺寸对比)
│ ├─ all_factor/ 16 ckpt,full-factor(对齐 pi0.5 协议)
│ └─ conflict_env/ 10 实验,双轴 env success(seed42)
├─ pi0/pairgrid_hard/ color_spatial(9 ckpt)+ color_size(3 ckpt),硬口径
├─ genie/conflict_env/ 10 实验,双轴 env success
└─ vlm_gemini-2.5-flash/ Gemini-2.5-flash 看视频判 factor_followed + FDR_summary.txt
simulation/
├─ eval_simulation/ ManiSkill 仿真代码(VerbObjectColor-v1 环境)
├─ Maniskill_gen_new/ 数据采集/环境定义(get_env_id_and_color)
└─ eval_simulation_ASSETS.tar.gz 二进制 3D 资源(解开覆盖回 eval_simulation/)
每个 ckpt 目录里:*_seed42.txt / _seed40.txt / _seed41.txt(逐 episode 明细)+ SUMMARY.txt(3-seed 均)。
checkpoint 权重(本仓库不含)
- pair-grid(f6/f12/f18):HF
yqi19/gr00t_public_pair_ckpt - all_factor(f50/n*):HF
yqi19/gr00t_all_factor_evaluation
评测协议(一句话)
- pairgrid 硬口径:同色(或同形)cube 干扰项,强制模型靠目标因子区分;task_difficulty=1.5,gpu-sim,150 步,replan=10,210 ep/seed,seeds 42/40/41。
- pairgrid 软口径(EASY):仅 color_size,1 干扰项 + ~2× 尺寸对比。
- all_factor:单指令全因子采样(pi0.5 对齐),200 ep,500 步,no_distractor=0.70。
- conflict env:双轴成功率(因子A / 因子B 各自命中)。
- VLM:Gemini-2.5-flash 看视频判模型跟了哪个因子,算 FDR。
每个实验的 target/distractor 尺寸、spatial 坐标等精确配置见本文件末尾「附:干扰项明细」。
关键结果(3-seed 均,各实验最强 ckpt)
| 实验 | 模型 | 最强 ckpt | 成功率 |
|---|---|---|---|
| color_spatial | pi0 | stair_f18 | 54.9% |
| color_spatial | GR00T | stair_f18 | 51.1% |
| spatial_size | GR00T | stair_f18 | 48.1% |
| all_factor | GR00T | Lrandom_f50_n400 | 47.0% |
| spatial_object | GR00T | Lrandom_f12 | 39.5% |
| color_size (EASY) | GR00T | Lrandom_f18 | 37.2% |
| verb_spatial | GR00T | stair_f18 | 35.5% |
| color_size (HARD) | GR00T | Lrandom_f18 | 34.9% |
| verb_size | GR00T | random_f12 | 24.2% |
结论: 任务难度 color_spatial > spatial_size > spatial_object > verb_spatial > color_size ≈ verb_size;
含 size 的任务对 VLA 最难(8–35%);空间类任务里 stair_f18 普遍最强;random_f18 多实验偏弱。
完整逐 ckpt 表见 result.md(或各 results/.../SUMMARY.txt)。
- Downloads last month
- 1,019