The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/harbor/harbor.py", line 171, in _split_generators
raise DataFilesNotFoundError("No task.toml or instruction.md files found")
datasets.exceptions.DataFilesNotFoundError: No task.toml or instruction.md files found
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.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.
MobileManipBench (MagicSim large-object tasks): results and progress
Updated 2026-11-02 06:38 UTC (generated, do not edit). 20 tasks × 3 robots (go2plus, stretch3, sawyer_fin) = 60 task–robot pairs; each pair has Easy / Medium / Hard + DR = 180 level cells.
怎么重建这个页面、监控进度的脚本(给别的 agent 看):整个文件夹 rebuild_kit/,从 00_START_HERE.md 开始。页面的版式照 yyoraa/mobilemanipbench-results。
Progress
| finished tasks (all 3 robots: Easy/Medium/Hard success, DR ≥ 5 successful draws, owner validated) | 0 / 20 |
| task–robot pairs finished (simulation stages done + owner validated) | 0 / 60 |
| task–robot pairs with all stages done, waiting for the owner | 20 / 60 |
| task–robot pairs with Easy, Medium and Hard all successful | 23 / 60 |
| task–robot pairs with DR ≥ 5 successful draws | 31 / 60 |
| level cells (Easy / Medium / Hard) with a real success | 91 / 180 |
A success is the environment's own full-success termination (no teleport, no weld, no scripted motion; NavMesh + DWB base, MagicSim command chain). A pair is finished only after the owner validated it.
There are no videos in this project: stage 3 is the domain-randomisation draws (object instance swap + object pose DR, the base must really move). (The raw numbers are in INDEX.md and status_table/.)
需要你看一下(卡住的格子)
规则:一个格子在当前阶段累计活跃满 2 小时还没过,agent 就不再自己反复调,把失败原因和证据交给你看,你看了告诉我们问题出在哪里,agent 同时去做下一个格子。下表列出已满 2 小时、仍有 kit 在跑、当前阶段还没过的格子和它最近一次失败的原因:
| Agent | Task | Robot | 当前阶段 | 累计活跃 | 这格跑过 | 最近一次失败的原因 |
|---|---|---|---|---|---|---|
| a02 | F006 | go2plus | 阶段一 Easy | 约 4.3 天 | 133 次 | run ended before the episode did (OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 MiB. GPU 0 has a total capacity of 23.52 GiB of which 2.31 MiB is |
| a01 | F004 | go2plus | 阶段一 Easy | 约 3.0 天 | 71 次 | walltime: running |
| a02 | F007 | stretch3 | 阶段一 Easy | 约 36.5 小时 | 44 次 | unknown (episode reset without success) |
现在谁在做什么
只列有 kit 正在运行的格子(一个 kit 就是一次仿真任务,由显卡上的进程判断)。"当前阶段"是表里第一个还没通过的阶段;"正在跑"是这个 kit 跑的难度(DR = 域随机化的一次抽样)。"这格已待"是从这个任务、这个机器人的第一份运行记录算起的时间,"累计活跃"是把相邻运行记录的间隔(单段最多算 1 小时)加起来的估计,不是 GPU 时间。 一个 agent 同时可以在几个格子上有 kit;没有 kit 的 agent 写"暂无 kit 在跑"(它可能在改代码或等待显存)。
| Agent | Task | Robot | 当前阶段 | 正在跑 | 这格的首次运行(美国中部) | 这格已待 | 累计活跃 | 最近一次运行记录 |
|---|---|---|---|---|---|---|---|---|
| a01 | F001 把移动白板推到展示位并转向观众 | stretch3 | 阶段三 DR | hard | 10-22 23:17 | 10.1 天 | 约 6.6 天 | 59 分钟前 |
| a01 | F004 把轻型拼桌推到另一张桌边 | go2plus | 阶段一 Easy | hard | 10-23 11:00 | 9.6 天 | 约 3.0 天 | 2.2 小时前 |
| a02 | F006 取一块长板,在两个支座之间搭桥 | go2plus | 阶段一 Easy | DR | 10-22 10:55 | 10.6 天 | 约 4.3 天 | 40 分钟前 |
| a02 | F006 取一块长板,在两个支座之间搭桥 | sawyer_fin | 阶段三 DR | DR | 10-22 06:33 | 10.8 天 | 约 3.9 天 | 0 分钟前 |
| a02 | F007 送大漏斗到接料位,再倒入三颗球 | stretch3 | 阶段一 Easy | DR | 10-22 06:46 | 10.8 天 | 约 36.5 小时 | 6.8 小时前 |
| a03 | F014 搬球压下跷跷板,让另一端抬起 | stretch3 | 阶段三 DR | DR | 11-01 11:52 | 12.8 小时 | 约 12.4 小时 | 1.4 小时前 |
| a04 | F019 记住人选的图案,再把双面大展牌转到正确面 | stretch3 | 阶段三 DR | hard | 10-30 10:18 | 2.6 天 | 约 4.9 小时 | 14 分钟前 |
每个格子用了多久
上图画的就是上面"现在谁在做什么"那张表里的数字:条长 = 这个格子的第一份运行记录到现在("这格已待"),括号里是"累计活跃"。柱子的颜色是它现在所处的阶段。 下图和下面的表是一个阶段从开始到完成用了多久,从 11-02 00:37(美国中部)开始记录,这之前完成的阶段没有记录;第一次被看到时还在阶段一的格子,从它的第一份运行记录算起,其余的从记录开始算起。 时间分辨率约 10 分钟,约每 10 分钟更新一次。"活跃"只算有运行记录的时间,agent 想问题、改代码的时间不在里面。

还没有格子完成阶段。统计刚开始,完成一个阶段就会在这里多一行。
总表
每个机器人下面有五列:**Easy、Medium、Hard、DR(阶段三,蓝底)、人工验证(紫底)**。阶段一 = Easy 成功;阶段二 = Medium 和 Hard 都成功;阶段三 = DR 至少 5 次成功;人工验证只由你来标。
| 符号 | 出现在哪 | 意思 |
|---|---|---|
| ✅ | Easy、Medium、Hard | 这个难度最近一次标准运行(不带 DR)环境判定完全成功 |
✅ s/N |
DR | N 次 DR 抽样里 s 次成功,s ≥ 5,阶段三完成 |
| ✅ | 人工 | 你已经验证通过,这一对(任务 × 机器人)才算完成 |
| ❌ | Easy、Medium、Hard | 跑过,但最近一次没成功 |
❌ s/N |
DR | 跑过 N 次抽样只成功 s 次(还不到 5 次) |
| · | 任意列 | 还没有 |
| 绿色加粗 ▶ a02 运行 · 已3.1小时 | 正在跑的那一列 | 这个 agent 的 kit 正在这个难度(或 DR)上跑;后面是这个格子已经待了多久(和"现在谁在做什么"表里一致) |
格子下面的小字是该结果出现的时间(美国中部时间,月-日 时:分)。
| Task | Agent | go2plus | stretch3 | sawyer_fin | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Easy | Medium | Hard | DR 阶段三 | 人工 验证 | Easy | Medium | Hard | DR 阶段三 | 人工 验证 | Easy | Medium | Hard | DR 阶段三 | 人工 验证 | ||
| F001 把移动白板推到展示位并转向观众 | a01 | ❌ 10-27 06:36 | ✅ 10-28 02:11 | ❌ 10-25 22:09 | ✅ 24/67 10-26 22:17 | · | ✅ 10-23 06:19 | ✅ 10-24 03:01 | ✅ 10-23 04:10 ▶ a01 运行 · 已10.1 天 | ✅ 122/192 10-23 07:35 | · | ✅ 10-22 15:42 | ✅ 10-22 17:34 | ✅ 10-22 19:21 | ✅ 10/10 10-23 02:29 | · |
| F002 沿屏风外缘移动,把两折屏风展开成L形 | a01 | ✅ 10-27 05:34 | ❌ 10-29 01:41 | ✅ 10-25 20:00 | ✅ 26/68 10-26 20:18 | · | ✅ 10-23 05:49 | ✅ 10-23 06:13 | ✅ 10-24 16:33 | ✅ 146/178 10-23 07:42 | · | ✅ 10-22 15:06 | ✅ 10-22 17:59 | ✅ 10-22 19:29 | ✅ 10/10 10-23 02:37 | · |
| F003 拉出长桌延伸板并停在终点 | a01 | ✅ 10-22 15:43 | ✅ 10-23 01:25 | ✅ 10-25 18:33 | ✅ 37/63 10-23 15:00 | · | ✅ 10-23 05:53 | ✅ 10-23 06:44 | ✅ 10-22 21:38 | ✅ 171/251 10-24 07:46 | · | ✅ 10-22 15:18 | ✅ 10-22 18:08 | ✅ 10-22 19:39 | ✅ 10/10 10-23 02:47 | · |
| F004 把轻型拼桌推到另一张桌边 | a01 | ❌ 10-29 01:29 | ✅ 10-29 02:31 | ❌ 10-25 19:44 ▶ a01 运行 · 已9.6 天 | ❌ 0/68 | · | ✅ 10-22 15:29 | ✅ 10-26 23:17 | ❌ 10-29 01:01 | ✅ 48/261 10-24 00:27 | · | ✅ 10-22 15:28 | ✅ 10-22 18:32 | ✅ 10-22 16:57 | ✅ 10/10 10-23 03:22 | · |
| F005 把宽箱盖滑到旁侧支架,露出箱口 | a01 | ✅ 10-29 00:53 | ✅ 10-29 02:44 | ❌ 10-26 19:10 | ✅ 10/67 10-28 18:58 | · | ✅ 10-22 15:48 | ✅ 10-22 18:14 | ✅ 10-23 04:57 | ✅ 223/249 10-23 07:58 | · | ✅ 10-22 15:48 | ✅ 10-22 18:42 | ✅ 10-23 03:36 | ✅ 10/10 10-23 03:27 | · |
| F006 取一块长板,在两个支座之间搭桥 | a02 | ❌ 10-26 17:58 | ❌ 10-26 07:41 | ❌ 10-26 18:59 | ❌ 0/111 ▶ a02 运行 · 已10.6 天 | · | ✅ 10-23 04:28 | ✅ 10-26 03:25 | ✅ 10-26 03:31 | ✅ 109/246 10-26 13:14 | · | ✅ 10-22 22:14 | ✅ 10-22 15:28 | ✅ 10-22 15:45 | ✅ 184/198 10-26 05:16 ▶ a02 运行 · 已10.8 天 | · |
| F007 送大漏斗到接料位,再倒入三颗球 | a02 | ❌ 11-01 10:42 | ❌ 10-26 09:17 | ❌ 10-25 04:57 | ❌ 0/5 | · | ❌ 11-01 01:17 | ❌ 11-01 01:53 | ❌ 11-01 02:02 | ❌ 0/14 ▶ a02 运行 · 已10.8 天 | · | ❌ 11-01 01:48 | ❌ 11-01 01:57 | ❌ 11-01 02:09 | ❌ 1/43 | · |
| F008 找到指定大画框并斜靠到展示架 | a02 | ❌ 11-01 11:46 | ❌ 10-26 09:26 | ❌ 10-25 05:37 | ❌ 0/11 | · | ❌ 11-01 02:27 | ❌ 11-01 02:55 | ❌ 11-01 03:23 | ❌ 0/17 | · | ❌ 11-01 02:14 | ❌ 11-01 02:19 | ❌ 11-01 02:26 | ❌ 0/44 | · |
| F009 把空料箱送到传送带末端接料 | a02 | ❌ 10-30 03:03 | ❌ 10-26 05:04 | ❌ 10-26 05:31 | ❌ 2/129 | · | ❌ 10-26 11:14 | ❌ 10-26 04:30 | ❌ 10-26 04:34 | ❌ 0/191 | · | ✅ 10-22 18:39 | ✅ 10-23 01:56 | ✅ 10-26 05:42 | ✅ 195/206 10-26 04:12 | · |
| F010 把展台上的大防尘布揭到后侧 | a02 | ❌ 11-01 12:14 | ❌ 10-25 21:55 | ❌ 10-25 22:42 | ❌ 0/5 | · | ❌ 11-01 03:34 | ❌ 11-01 03:37 | ❌ 11-01 04:07 | ❌ 0/17 | · | ❌ 11-01 02:29 | ❌ 11-01 02:54 | ❌ 11-01 02:58 | ❌ 0/51 | · |
| F011 补上多米诺缺口,再推倒短链 | a03 | ✅ 11-01 07:38 | ❌ 10-29 03:37 | ❌ 10-23 11:09 | ❌ 0/37 | · | ✅ 11-01 11:36 | ❌ 10-29 01:32 | ❌ 10-29 01:40 | ✅ 48/131 11-01 13:15 | · | ✅ 11-01 13:53 | ❌ 10-31 01:07 | ❌ 10-30 10:00 | ✅ 38/73 10-31 01:56 | · |
| F012 送来接收盒,再拔挡片让球落入 | a03 | ❌ 10-28 20:53 | · | · | ❌ 0/3 | · | ✅ 10-28 18:53 | ❌ 10-29 11:45 | ❌ 10-23 09:28 | ❌ 3/24 | · | ❌ 10-29 00:45 | · | · | ❌ 0/9 | · |
| F013 舀起几颗玩具豆,跨台倒进碗里 | a03 | ❌ 10-29 00:31 | · | · | ❌ 0/5 | · | ❌ 10-28 21:30 | · | · | ❌ 0/10 | · | ❌ 10-29 00:40 | · | · | ❌ 0/7 | · |
| F014 搬球压下跷跷板,让另一端抬起 | a03 | ✅ 11-01 01:04 | · | ❌ 10-30 21:58 | ✅ 25/42 11-01 03:55 | · | ✅ 11-01 13:20 | ✅ 11-01 12:58 | ✅ 11-01 14:35 | ✅ 192/266 11-01 13:58 ▶ a03 运行 · 已12.8 小时 | · | ✅ 10-31 01:32 | ✅ 11-01 19:18 | ✅ 11-01 08:16 | ✅ 68/68 10-31 01:50 | · |
| F015 取大圈,圈住两颗球并推到收集区 | a03 | ❌ 11-01 16:18 | · | · | ❌ 0/12 | · | ❌ 11-01 13:28 | · | · | ✅ 9/67 11-01 17:19 | · | ✅ 11-01 00:07 | ✅ 11-01 14:19 | ✅ 10-31 01:40 | ✅ 43/79 10-31 02:57 | · |
| F016 把落地风扇移到指定位置并朝向人 | a04 | ❌ 10-31 23:43 | ❌ 11-01 23:26 | ❌ 11-02 00:19 | ❌ 3/5 | · | ✅ 10-30 08:54 | ❌ 11-01 22:21 | ✅ 11-01 22:27 | ✅ 5/5 11-01 07:21 | · | ✅ 11-02 00:01 | ✅ 11-01 23:31 | ✅ 10-31 22:15 | ❌ 3/5 | · |
| F017 拉空行李箱,按人的停/继续手势送达 | a04 | ✅ 10-28 05:57 | ✅ 11-01 03:44 | ✅ 11-01 04:15 | ❌ 3/5 | · | ✅ 10-30 12:12 | ✅ 11-01 22:52 | ✅ 11-01 23:25 | ✅ 5/5 10-30 17:49 | · | ✅ 10-30 11:19 | ✅ 10-30 11:34 | ✅ 10-30 11:39 | ✅ 5/5 11-01 09:53 | · |
| F018 携宽空托盘让人先通过,再送到收纳架 | a04 | ❌ 10-31 23:19 | ❌ 11-01 22:16 | ❌ 11-01 22:46 | ❌ 0/5 | · | ✅ 10-30 09:23 | ✅ 11-01 23:51 | ✅ 11-02 00:10 | ✅ 5/5 11-01 08:08 | · | ✅ 11-02 00:25 | ✅ 11-02 00:19 | ❌ 11-01 04:30 | ❌ 1/5 | · |
| F019 记住人选的图案,再把双面大展牌转到正确面 | a04 | ✅ 10-31 23:04 | ❌ 11-01 01:03 | ❌ 11-01 01:58 | ❌ 0/5 | · | ✅ 10-30 10:18 | ✅ 11-02 00:24 | ✅ 11-01 05:48 ▶ a04 运行 · 已2.6 天 | ❌ 4/5 | · | ✅ 10-30 10:14 | ✅ 10-31 23:07 | ❌ 11-01 05:01 | ✅ 5/5 11-01 08:41 | · |
| F020 绕到大转盘另一侧,把指定展品转到人面前 | a04 | ❌ 10-31 22:11 | ❌ 11-01 00:47 | ✅ 11-01 01:17 | ❌ 2/5 | · | ✅ 10-30 11:28 | ✅ 10-30 11:47 | ✅ 11-01 05:59 | ✅ 5/5 11-01 09:18 | · | ✅ 10-30 10:43 | ❌ 10-31 23:31 | ❌ 11-01 05:16 | ✅ 5/5 11-01 09:30 | · |
Where things are
INDEX.md/index.csv: one row per task × robot with the result, the agent, Easy/Medium/Hard, DR, human validation, run count and the last failure reason.by_agent/aNN.md: the agent's five tasks on the three robots. Agent a01 has F001–F005, a02 F006–F010, a03 F011–F015, a04 F016–F020.cells/<task>_<title>_<agent>/<robot>/STATUS.md: that pair's results in words (all 60 folders always exist).rebuild_kit/: everything needed to rebuild this page:00_START_HERE.md,FRONT_PAGE_SPEC.md,HOW_TO_REBUILD.md,scripts/(the table builder, the page generator, the chart drawer).status_table/: the live progress table (csv, json), the input of the page generator.
About the records
The simulation records themselves (per-step json, camera frames) are not in this repo; they are on the owner's private Hugging Face backup. This page only reports the results of the runs. A path never changes when a pair is validated: the validation state lives in INDEX.md.
- Downloads last month
- 123