task_id stringlengths 3 3 | variant stringclasses 3
values | title stringlengths 4 12 | task_json stringlengths 21 21 | prompt stringlengths 21 21 | text_rule stringlengths 24 24 | real dict | sim dict | sim_status stringclasses 1
value | prompt_text stringlengths 166 244 | text_context stringlengths 21 52 |
|---|---|---|---|---|---|---|---|---|---|---|
E01 | A | 套餐订单履约 | tasks/E01/A/task.json | tasks/E01/A/prompt.md | tasks/E01/A/text_rule.md | {
"path": "tasks/E01/A/real.mp4",
"sha256": "e9e3749e9e0ebb47912608c5e47a3fda809ecf7e7e4303c5fa04268522d20027",
"bytes": 9363174,
"duration_seconds": 12.4,
"frame_count": 372,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9d0e11c090... | {
"path": "tasks/E01/A/sim.mp4",
"sha256": "9becc46b832d34f48ee0f9c28547cb6e1a9d2faaa655a8cd55a6756dcaddd05a",
"bytes": 1281116,
"duration_seconds": 45.08,
"frame_count": 1127,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "08d0f20e48b... | available | # E01A · 套餐订单履约
## 执行任务目标
根据台面订单,为标记的取餐位准备一份套餐并装箱送达。食物组合按视频中的选配方式;订单只指定取餐位和包装,不指定组合。非套餐食物留在备料区,餐盘内物品不得掉出,交付后工作区不留包装垃圾。
## 原始局部示范任务
按示范中的食物选择组合,把相应食物装入餐盘,其余留在桌上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E01A 文字规则
汉堡和薯条装盘,饮料留外
|
E01 | B | 套餐订单履约 | tasks/E01/B/task.json | tasks/E01/B/prompt.md | tasks/E01/B/text_rule.md | {
"path": "tasks/E01/B/real.mp4",
"sha256": "b80bd47179002610b09c2787cd3675332b33aa3d3fc56c572474809c3d09bdfc",
"bytes": 7424703,
"duration_seconds": 9.766667,
"frame_count": 293,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "4d60f0... | {
"path": "tasks/E01/B/sim.mp4",
"sha256": "7776e3016ed9719c24f322fc2733de8f6fc62cd223e548502f3b8763f7e51ef1",
"bytes": 1235005,
"duration_seconds": 45.4,
"frame_count": 1135,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "a303ea1f87d2... | available | # E01B · 套餐订单履约
## 执行任务目标
根据台面订单,为标记的取餐位准备一份套餐并装箱送达。食物组合按视频中的选配方式;订单只指定取餐位和包装,不指定组合。非套餐食物留在备料区,餐盘内物品不得掉出,交付后工作区不留包装垃圾。
## 原始局部示范任务
按示范中的食物选择组合,把相应食物装入餐盘,其余留在桌上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E01B 文字规则
汉堡和饮料装盘,薯条留外
|
E01 | C | 套餐订单履约 | tasks/E01/C/task.json | tasks/E01/C/prompt.md | tasks/E01/C/text_rule.md | {
"path": "tasks/E01/C/real.mp4",
"sha256": "605e5aae8c6eabeace8e03b4f9e245a4464f9651f79cd37ace319e61e9de511a",
"bytes": 7273662,
"duration_seconds": 9.6,
"frame_count": 288,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5765e7c55a3... | {
"path": "tasks/E01/C/sim.mp4",
"sha256": "5d0aa092cb034b9e63ba6a6e13192e47b585f2c40f068b1031c5a01cb1a7bb4c",
"bytes": 1276227,
"duration_seconds": 48.4,
"frame_count": 1210,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "48a402f3ced7... | available | # E01C · 套餐订单履约
## 执行任务目标
根据台面订单,为标记的取餐位准备一份套餐并装箱送达。食物组合按视频中的选配方式;订单只指定取餐位和包装,不指定组合。非套餐食物留在备料区,餐盘内物品不得掉出,交付后工作区不留包装垃圾。
## 原始局部示范任务
按示范中的食物选择组合,把相应食物装入餐盘,其余留在桌上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E01C 文字规则
薯条和饮料装盘,汉堡留外
|
E02 | A | 三色块叠放 | tasks/E02/A/task.json | tasks/E02/A/prompt.md | tasks/E02/A/text_rule.md | {
"path": "tasks/E02/A/real.mp4",
"sha256": "6a9267465873da3b6914085b235b372c9af81581891d626ab75a96f9c0e1fce5",
"bytes": 7565483,
"duration_seconds": 12.033333,
"frame_count": 361,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "f7348... | {
"path": "tasks/E02/A/sim.mp4",
"sha256": "24f31eab6b8835be2d82384e513ac6435e65d9e90e9b14447c334c50ebf8fb88",
"bytes": 831077,
"duration_seconds": 34.08,
"frame_count": 852,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "8b5b00566c136... | available | # E02A · 三色块叠放
## 执行任务目标
按示范中的颜色层序,把三个色块叠成稳定的一摞。
## 原始局部示范任务
按示范中的颜色层序,把三个色块叠成稳定的一摞。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E02A 文字规则
自下而上红、绿、蓝
|
E02 | B | 三色块叠放 | tasks/E02/B/task.json | tasks/E02/B/prompt.md | tasks/E02/B/text_rule.md | {
"path": "tasks/E02/B/real.mp4",
"sha256": "96ffc7758f062cc66dfbe28844e0beefb108f5efeb2d2fcb8a5b09feeccd56b4",
"bytes": 8636332,
"duration_seconds": 13.5,
"frame_count": 405,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "25adcfc599... | {
"path": "tasks/E02/B/sim.mp4",
"sha256": "4ac2b7866a1f07f8a6421eff4d0465bc36b460c3fba97bf40776e0af1f374605",
"bytes": 818823,
"duration_seconds": 33.72,
"frame_count": 843,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ea4ca4915cd75... | available | # E02B · 三色块叠放
## 执行任务目标
按示范中的颜色层序,把三个色块叠成稳定的一摞。
## 原始局部示范任务
按示范中的颜色层序,把三个色块叠成稳定的一摞。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E02B 文字规则
自下而上绿、蓝、红
|
E02 | C | 三色块叠放 | tasks/E02/C/task.json | tasks/E02/C/prompt.md | tasks/E02/C/text_rule.md | {
"path": "tasks/E02/C/real.mp4",
"sha256": "3ae380e5738752b97cee551f8c969c498a0d96abf22c95b6e05a2f593d57567e",
"bytes": 7992854,
"duration_seconds": 12.666667,
"frame_count": 380,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "bc750... | {
"path": "tasks/E02/C/sim.mp4",
"sha256": "a4609d953c35fe22d66c5abb70e873a79b132142b40223fd8d820d7d6308739d",
"bytes": 810344,
"duration_seconds": 33.08,
"frame_count": 827,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ba272e0b1edcb... | available | # E02C · 三色块叠放
## 执行任务目标
按示范中的颜色层序,把三个色块叠成稳定的一摞。
## 原始局部示范任务
按示范中的颜色层序,把三个色块叠成稳定的一摞。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E02C 文字规则
自下而上蓝、红、绿
|
E03 | A | 三色杯套叠 | tasks/E03/A/task.json | tasks/E03/A/prompt.md | tasks/E03/A/text_rule.md | {
"path": "tasks/E03/A/real.mp4",
"sha256": "42a7350c20f94bd37835fd840159c8c3facf42ee1bdd3c2fd1afec383072ac12",
"bytes": 8502641,
"duration_seconds": 13.333333,
"frame_count": 400,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5b2f7... | {
"path": "tasks/E03/A/sim.mp4",
"sha256": "bb806b0dba7f0424f359a40d6c9db4b32aba93486ba8abb5be927f25c6e6dac2",
"bytes": 2332875,
"duration_seconds": 113.08,
"frame_count": 2827,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "bae7ccfe43... | available | # E03A · 三色杯套叠
## 执行任务目标
按示范中的颜色内外次序,把三个杯子套在一起。
## 原始局部示范任务
按示范中的颜色内外次序,把三个杯子套在一起。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E03A 文字规则
外层到内层:红、绿、蓝
|
E03 | B | 三色杯套叠 | tasks/E03/B/task.json | tasks/E03/B/prompt.md | tasks/E03/B/text_rule.md | {
"path": "tasks/E03/B/real.mp4",
"sha256": "b88ffc42a5411439a910c907594c36fe821631d7528d24568dca186a5cb2197f",
"bytes": 7046018,
"duration_seconds": 11.4,
"frame_count": 342,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "d2384d0500... | {
"path": "tasks/E03/B/sim.mp4",
"sha256": "3c407ea3f780beab5c5cca88a57bbadd45abb325d8017c662a181e52056ad610",
"bytes": 2287458,
"duration_seconds": 110.44,
"frame_count": 2761,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "1fc09b361d... | available | # E03B · 三色杯套叠
## 执行任务目标
按示范中的颜色内外次序,把三个杯子套在一起。
## 原始局部示范任务
按示范中的颜色内外次序,把三个杯子套在一起。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E03B 文字规则
外层到内层:绿、蓝、红
|
E03 | C | 三色杯套叠 | tasks/E03/C/task.json | tasks/E03/C/prompt.md | tasks/E03/C/text_rule.md | {
"path": "tasks/E03/C/real.mp4",
"sha256": "8aabb465563f8bf8918a7a9732e5226a9f921d876060306573eabf1f8894f8c4",
"bytes": 7971112,
"duration_seconds": 12.633333,
"frame_count": 379,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "8178c... | {
"path": "tasks/E03/C/sim.mp4",
"sha256": "9b937ca22dca437fed10a65150504a5e438ab08ff07b1496d75ee965ff5bdc4e",
"bytes": 2227194,
"duration_seconds": 108.04,
"frame_count": 2701,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "b69898d288... | available | # E03C · 三色杯套叠
## 执行任务目标
按示范中的颜色内外次序,把三个杯子套在一起。
## 原始局部示范任务
按示范中的颜色内外次序,把三个杯子套在一起。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E03C 文字规则
外层到内层:蓝、红、绿
|
E04 | A | 彩杯与图形配对 | tasks/E04/A/task.json | tasks/E04/A/prompt.md | tasks/E04/A/text_rule.md | {
"path": "tasks/E04/A/real.mp4",
"sha256": "e3a7a20ebf30e5751340cfc1c0be1db6cbf8246687add559771f8ca22b170c4f",
"bytes": 6583027,
"duration_seconds": 9.8,
"frame_count": 294,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9eef7940163... | {
"path": "tasks/E04/A/sim.mp4",
"sha256": "9495171e67d06888e35969facb66d5087e0632ea72832bc41464a1beebf9c024",
"bytes": 2420521,
"duration_seconds": 107.28,
"frame_count": 2682,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c4f0db080a... | available | # E04A · 彩杯与图形配对
## 执行任务目标
按示范中杯子颜色与图形的对应关系,把三个杯子摆到相应标记上。
## 原始局部示范任务
按示范中杯子颜色与图形的对应关系,把三个杯子摆到相应标记上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E04A 文字规则
绿杯→三角;蓝杯→圆;红杯→方
|
E04 | B | 彩杯与图形配对 | tasks/E04/B/task.json | tasks/E04/B/prompt.md | tasks/E04/B/text_rule.md | {
"path": "tasks/E04/B/real.mp4",
"sha256": "c4d5571a64c78aa0a53728c184c35d4ba0672127ee2b3843d32841e2693f1b57",
"bytes": 6342801,
"duration_seconds": 9.466667,
"frame_count": 284,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "af44e5... | {
"path": "tasks/E04/B/sim.mp4",
"sha256": "7b42848a327f05f57b98003601802af7b534ecf58aebf54b3a9fcee14aa96fc0",
"bytes": 1432236,
"duration_seconds": 59.4,
"frame_count": 1485,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "a7a9492b1bd1... | available | # E04B · 彩杯与图形配对
## 执行任务目标
按示范中杯子颜色与图形的对应关系,把三个杯子摆到相应标记上。
## 原始局部示范任务
按示范中杯子颜色与图形的对应关系,把三个杯子摆到相应标记上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E04B 文字规则
绿杯→方;蓝杯→圆;红杯→三角
|
E04 | C | 彩杯与图形配对 | tasks/E04/C/task.json | tasks/E04/C/prompt.md | tasks/E04/C/text_rule.md | {
"path": "tasks/E04/C/real.mp4",
"sha256": "77557ab280f0e64a9d58f524ec8ab9f0f7826c30420ada2bfd7e6120fff61145",
"bytes": 6262298,
"duration_seconds": 9.366667,
"frame_count": 281,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "981f3f... | {
"path": "tasks/E04/C/sim.mp4",
"sha256": "08c908702cb66801d44f752ca39546da9a64bd797396e868636ec6a75efad516",
"bytes": 1503586,
"duration_seconds": 60.8,
"frame_count": 1520,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "e75f0185fc29... | available | # E04C · 彩杯与图形配对
## 执行任务目标
按示范中杯子颜色与图形的对应关系,把三个杯子摆到相应标记上。
## 原始局部示范任务
按示范中杯子颜色与图形的对应关系,把三个杯子摆到相应标记上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E04C 文字规则
绿杯→圆;蓝杯→三角;红杯→方
|
E05 | A | 形状块与图形配对 | tasks/E05/A/task.json | tasks/E05/A/prompt.md | tasks/E05/A/text_rule.md | {
"path": "tasks/E05/A/real.mp4",
"sha256": "ec496a155ce4de38d033824db9eaea82ede633fe41aa15eb4037e38ea92deee5",
"bytes": 6209134,
"duration_seconds": 9.633333,
"frame_count": 289,
"fps": "30/1",
"width": 1280,
"height": 720,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5764d92... | {
"path": "tasks/E05/A/sim.mp4",
"sha256": "c4b622afbb4a3b604dada5d281792e784c71e837174a38cbd9599cc94002105f",
"bytes": 1541373,
"duration_seconds": 59.96,
"frame_count": 1499,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "172694562f2... | available | # E05A · 形状块与图形配对
## 执行任务目标
按示范中物体形状与图形的对应关系,摆放全部三个物体。
## 原始局部示范任务
按示范中物体形状与图形的对应关系,摆放全部三个物体。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E05A 文字规则
方块→方形;球→圆形;三角块→三角形
|
E05 | B | 形状块与图形配对 | tasks/E05/B/task.json | tasks/E05/B/prompt.md | tasks/E05/B/text_rule.md | {
"path": "tasks/E05/B/real.mp4",
"sha256": "fc5d0a9964ced96641674afd0159daf542ea90d631a77ebb5a5e0367630cc402",
"bytes": 7208557,
"duration_seconds": 10.566667,
"frame_count": 317,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "49b26... | {
"path": "tasks/E05/B/sim.mp4",
"sha256": "9dd8c77c60f47d25fea44b7e46efc827161bdfbb79f778265a1f87b9b640d600",
"bytes": 1692088,
"duration_seconds": 65.52,
"frame_count": 1638,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "93fd0db564b... | available | # E05B · 形状块与图形配对
## 执行任务目标
按示范中物体形状与图形的对应关系,摆放全部三个物体。
## 原始局部示范任务
按示范中物体形状与图形的对应关系,摆放全部三个物体。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E05B 文字规则
方块→圆;球→三角;三角→方
|
E05 | C | 形状块与图形配对 | tasks/E05/C/task.json | tasks/E05/C/prompt.md | tasks/E05/C/text_rule.md | {
"path": "tasks/E05/C/real.mp4",
"sha256": "98f6387477cae0dec7f8fda12aad4b64b4d7a5a579fb50e0394543f6f003418b",
"bytes": 6983834,
"duration_seconds": 10.333333,
"frame_count": 310,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ed6ab... | {
"path": "tasks/E05/C/sim.mp4",
"sha256": "c2ef180a8f903744309a3ab2b7c1346f7773e1c93dee26fdab91170fc0efce84",
"bytes": 1523364,
"duration_seconds": 59.6,
"frame_count": 1490,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5c9bec3094b9... | available | # E05C · 形状块与图形配对
## 执行任务目标
按示范中物体形状与图形的对应关系,摆放全部三个物体。
## 原始局部示范任务
按示范中物体形状与图形的对应关系,摆放全部三个物体。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E05C 文字规则
方块→三角;球→方;三角→圆
|
E06 | A | 蛋托孔位布局 | tasks/E06/A/task.json | tasks/E06/A/prompt.md | tasks/E06/A/text_rule.md | {
"path": "tasks/E06/A/real.mp4",
"sha256": "1314469391881e2e884993228e6f1a1bffc495c45fb0711d296a99de6e54b8fa",
"bytes": 4811600,
"duration_seconds": 7.933333,
"frame_count": 238,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "22d0fc... | {
"path": "tasks/E06/A/sim.mp4",
"sha256": "242003d5e1e494dafec58f691651e449c41036d3beadb454f584ae6e261b415b",
"bytes": 611014,
"duration_seconds": 28.32,
"frame_count": 708,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9b592b9c495ef... | available | # E06A · 蛋托孔位布局
## 执行任务目标
按示范的孔位布局,把三枚球放入六孔托盘。左右以你面对托盘的方向为准。
## 原始局部示范任务
按示范的孔位布局,把三枚球放入六孔托盘。左右以你面对托盘的方向为准。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E06A 文字规则
填满左列三个孔
|
E06 | B | 蛋托孔位布局 | tasks/E06/B/task.json | tasks/E06/B/prompt.md | tasks/E06/B/text_rule.md | {
"path": "tasks/E06/B/real.mp4",
"sha256": "b2303edd53238427fb136546e523724eb01ec4d5bbe1000928d8cc2ef6d5ada9",
"bytes": 3922143,
"duration_seconds": 7.133333,
"frame_count": 214,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "1494a8... | {
"path": "tasks/E06/B/sim.mp4",
"sha256": "2e6727426afe572b54396f55752d684de12414a8d84e5d11e6aa9e3dd1aeb68b",
"bytes": 631041,
"duration_seconds": 29.64,
"frame_count": 741,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "98b972c947d67... | available | # E06B · 蛋托孔位布局
## 执行任务目标
按示范的孔位布局,把三枚球放入六孔托盘。左右以你面对托盘的方向为准。
## 原始局部示范任务
按示范的孔位布局,把三枚球放入六孔托盘。左右以你面对托盘的方向为准。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E06B 文字规则
填满右列三个孔
|
E06 | C | 蛋托孔位布局 | tasks/E06/C/task.json | tasks/E06/C/prompt.md | tasks/E06/C/text_rule.md | {
"path": "tasks/E06/C/real.mp4",
"sha256": "d7bd684145139127ee996c8ec75eb44e89bf85ccefe2cb2c91f8b3d714e20956",
"bytes": 4687501,
"duration_seconds": 8.1,
"frame_count": 243,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "7a3a6ee332a... | {
"path": "tasks/E06/C/sim.mp4",
"sha256": "c34f55077271c55c19f11477c017f6dfb180c7ad120a593cdf593540a536ab09",
"bytes": 613473,
"duration_seconds": 28.64,
"frame_count": 716,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "85cf6659bf0db... | available | # E06C · 蛋托孔位布局
## 执行任务目标
按示范的孔位布局,把三枚球放入六孔托盘。左右以你面对托盘的方向为准。
## 原始局部示范任务
按示范的孔位布局,把三枚球放入六孔托盘。左右以你面对托盘的方向为准。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E06C 文字规则
左列两端、右列中间各放一球
|
E07 | A | 五块尺寸排序 | tasks/E07/A/task.json | tasks/E07/A/prompt.md | tasks/E07/A/text_rule.md | {
"path": "tasks/E07/A/real.mp4",
"sha256": "6487a9115a68effedade7bb352339313921fcca465fa4234861ed9ffdb843cf5",
"bytes": 14012704,
"duration_seconds": 20.966667,
"frame_count": 629,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ec9d... | {
"path": "tasks/E07/A/sim.mp4",
"sha256": "00ef0887a1e1bdb36f36d86c52dece961d5b5023dff435fb4ba4cf910b2a2bf2",
"bytes": 1354707,
"duration_seconds": 42.84,
"frame_count": 1071,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "59c290db48c... | available | # E07A · 五块尺寸排序
## 执行任务目标
按照示范中的尺寸排序,把五块积木沿远到近方向排成一列。
## 原始局部示范任务
按照示范中的尺寸排序,把五块积木沿远到近方向排成一列。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E07A 文字规则
从远到近:最小、次小、中等、次大、最大
|
E07 | B | 五块尺寸排序 | tasks/E07/B/task.json | tasks/E07/B/prompt.md | tasks/E07/B/text_rule.md | {
"path": "tasks/E07/B/real.mp4",
"sha256": "8c8248767a1eecf58a0b3c6d4d20325a09b2436c319f0528e3f6390dc0c05f71",
"bytes": 12354402,
"duration_seconds": 18.433333,
"frame_count": 553,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3a49... | {
"path": "tasks/E07/B/sim.mp4",
"sha256": "98afe824820ccfcec2a7443b948f13e2d4566b7d7f8973fac4133553509f62a3",
"bytes": 1330426,
"duration_seconds": 42.48,
"frame_count": 1062,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "7f65028be63... | available | # E07B · 五块尺寸排序
## 执行任务目标
按照示范中的尺寸排序,把五块积木沿远到近方向排成一列。
## 原始局部示范任务
按照示范中的尺寸排序,把五块积木沿远到近方向排成一列。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E07B 文字规则
从远到近:最大、次大、中等、次小、最小
|
E07 | C | 五块尺寸排序 | tasks/E07/C/task.json | tasks/E07/C/prompt.md | tasks/E07/C/text_rule.md | {
"path": "tasks/E07/C/real.mp4",
"sha256": "23874b862470a336afa2704042c009bd17fc37a3ece4bba28b065f08350ef176",
"bytes": 9494228,
"duration_seconds": 15.266667,
"frame_count": 458,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "fc05d... | {
"path": "tasks/E07/C/sim.mp4",
"sha256": "b0f3ee47e2288186cfebc53bfb67697a0a184d58866ef397a2f30431703932df",
"bytes": 1368012,
"duration_seconds": 43.12,
"frame_count": 1078,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "84bad6c9591... | available | # E07C · 五块尺寸排序
## 执行任务目标
按照示范中的尺寸排序,把五块积木沿远到近方向排成一列。
## 原始局部示范任务
按照示范中的尺寸排序,把五块积木沿远到近方向排成一列。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E07C 文字规则
从远到近:中等、次小、最小、次大、最大
|
E08 | A | 鞋盒内朝向摆放 | tasks/E08/A/task.json | tasks/E08/A/prompt.md | tasks/E08/A/text_rule.md | {
"path": "tasks/E08/A/real.mp4",
"sha256": "634db3bccdc3eabef3f6852d8cb1b64cfece895b187a686c0fb104edc4383a64",
"bytes": 10737306,
"duration_seconds": 14.133333,
"frame_count": 424,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "cc2f... | {
"path": "tasks/E08/A/sim.mp4",
"sha256": "f7030cf3835db16c87fe8abe9e4ae85a2ac2544246956e8e4f4f8ba87575b533",
"bytes": 1558045,
"duration_seconds": 60.2,
"frame_count": 1505,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "4db989e0c4ea... | available | # E08A · 鞋盒内朝向摆放
## 执行任务目标
按示范中鞋尖与鞋盒标签端的朝向关系,把两只鞋放入鞋盒。
## 原始局部示范任务
按示范中鞋尖与鞋盒标签端的朝向关系,把两只鞋放入鞋盒。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E08A 文字规则
两只鞋尖均朝鞋盒标签端
|
E08 | B | 鞋盒内朝向摆放 | tasks/E08/B/task.json | tasks/E08/B/prompt.md | tasks/E08/B/text_rule.md | {
"path": "tasks/E08/B/real.mp4",
"sha256": "a0f0ca1a1bbcf2c6c1294e90cade15fe977f9e9c9703fc9cb5977675204b7af6",
"bytes": 7990216,
"duration_seconds": 10.5,
"frame_count": 315,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "df49a28670... | {
"path": "tasks/E08/B/sim.mp4",
"sha256": "297e83e822ff5908813d32128b72a0417638b5f53338a944ddef3c402c553d6e",
"bytes": 1069232,
"duration_seconds": 40.04,
"frame_count": 1001,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3e918618171... | available | # E08B · 鞋盒内朝向摆放
## 执行任务目标
按示范中鞋尖与鞋盒标签端的朝向关系,把两只鞋放入鞋盒。
## 原始局部示范任务
按示范中鞋尖与鞋盒标签端的朝向关系,把两只鞋放入鞋盒。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E08B 文字规则
两只鞋跟均朝标签端
|
E08 | C | 鞋盒内朝向摆放 | tasks/E08/C/task.json | tasks/E08/C/prompt.md | tasks/E08/C/text_rule.md | {
"path": "tasks/E08/C/real.mp4",
"sha256": "e29cd547dc299906ed0933e35783cba0c4a3a54078bdc765ef046a4088479c74",
"bytes": 8401918,
"duration_seconds": 11.066667,
"frame_count": 332,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "bec64... | {
"path": "tasks/E08/C/sim.mp4",
"sha256": "e403aab6a44f83fabbddd3a654c6bdba8485013a4d2bffa00e73cef452bdca64",
"bytes": 1263766,
"duration_seconds": 49.4,
"frame_count": 1235,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "f66ec363e0ef... | available | # E08C · 鞋盒内朝向摆放
## 执行任务目标
按示范中鞋尖与鞋盒标签端的朝向关系,把两只鞋放入鞋盒。
## 原始局部示范任务
按示范中鞋尖与鞋盒标签端的朝向关系,把两只鞋放入鞋盒。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E08C 文字规则
左脚鞋尖朝标签端,右脚鞋跟朝标签端
|
E09 | A | 分类教具配套出库 | tasks/E09/A/task.json | tasks/E09/A/prompt.md | tasks/E09/A/text_rule.md | {
"path": "tasks/E09/A/real.mp4",
"sha256": "ca2acd7c5515b8f3310dc7c9b2b055015268620ade76535d8b911f4d2bf5247d",
"bytes": 17822479,
"duration_seconds": 24.866667,
"frame_count": 746,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3dfa... | {
"path": "tasks/E09/A/sim.mp4",
"sha256": "5c0f2042ee718d1180016988c35bef34e86c4aebda7189028fc05f3b9720368a",
"bytes": 3840242,
"duration_seconds": 134,
"frame_count": 3350,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "8aeb3a987fe69... | available | # E09A · 分类教具配套出库
## 执行任务目标
为本次课堂准备完整三包教具,分组对应关系按视频。每包有三件且放入独立透明盒,盒内不得混组;将三盒装入订单指定的周转箱并送到对应领取区,备用件留在原架。
## 原始局部示范任务
按示范的颜色与形状分组关系,将九件物体分成三组,每组三件。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E09A 文字规则
同色一组:红、绿、蓝各三件
|
E09 | B | 分类教具配套出库 | tasks/E09/B/task.json | tasks/E09/B/prompt.md | tasks/E09/B/text_rule.md | {
"path": "tasks/E09/B/real.mp4",
"sha256": "e76e16df82adfde57cd31d69da628c708a4d206622a48ddb49d2c5d2a86f5e11",
"bytes": 14746624,
"duration_seconds": 21.5,
"frame_count": 645,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9e725fccb... | {
"path": "tasks/E09/B/sim.mp4",
"sha256": "ec5854a549a71e602ea1b1381125f66eed51c9eeef486fced64f7f09b11afc9b",
"bytes": 3844239,
"duration_seconds": 140.56,
"frame_count": 3514,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "d0de1212d1... | available | # E09B · 分类教具配套出库
## 执行任务目标
为本次课堂准备完整三包教具,分组对应关系按视频。每包有三件且放入独立透明盒,盒内不得混组;将三盒装入订单指定的周转箱并送到对应领取区,备用件留在原架。
## 原始局部示范任务
按示范的颜色与形状分组关系,将九件物体分成三组,每组三件。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E09B 文字规则
同形一组:方、三角、圆各三件
|
E09 | C | 分类教具配套出库 | tasks/E09/C/task.json | tasks/E09/C/prompt.md | tasks/E09/C/text_rule.md | {
"path": "tasks/E09/C/real.mp4",
"sha256": "e857abb045064ae9c435bcc18f2897ffa203c216d2604b2a02e7c2ad1d76cb73",
"bytes": 11270917,
"duration_seconds": 16.733333,
"frame_count": 502,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "e89f... | {
"path": "tasks/E09/C/sim.mp4",
"sha256": "dbed6406306f3c046073112cab27cb0ebb5353eb9496514711b412182f046e64",
"bytes": 3961760,
"duration_seconds": 137.24,
"frame_count": 3431,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "8a2e6cbd4e... | available | # E09C · 分类教具配套出库
## 执行任务目标
为本次课堂准备完整三包教具,分组对应关系按视频。每包有三件且放入独立透明盒,盒内不得混组;将三盒装入订单指定的周转箱并送到对应领取区,备用件留在原架。
## 原始局部示范任务
按示范的颜色与形状分组关系,将九件物体分成三组,每组三件。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E09C 文字规则
红圆+绿三角+蓝方;红三角+绿方+蓝圆;红方+绿圆+蓝三角
|
E10 | A | 按属性选择并套柱 | tasks/E10/A/task.json | tasks/E10/A/prompt.md | tasks/E10/A/text_rule.md | {
"path": "tasks/E10/A/real.mp4",
"sha256": "7f94389e2bd9e651bb71f1c5595975ec2d4fc938416652a5d41225f921444c1f",
"bytes": 10644581,
"duration_seconds": 15.233333,
"frame_count": 457,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "add0... | {
"path": "tasks/E10/A/sim.mp4",
"sha256": "290856ad5114de65d6c8d272e8411f2838a7968b620afec5649dc9183209edaf",
"bytes": 1971049,
"duration_seconds": 90,
"frame_count": 2250,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "37757941681d3f... | available | # E10A · 按属性选择并套柱
## 执行任务目标
按示范的颜色、形状选择方式,把符合条件的零件套到匹配的立柱上,其余留在桌上。
## 原始局部示范任务
按示范的颜色、形状选择方式,把符合条件的零件套到匹配的立柱上,其余留在桌上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E10A 文字规则
套入所有红色件,共3件
|
E10 | B | 按属性选择并套柱 | tasks/E10/B/task.json | tasks/E10/B/prompt.md | tasks/E10/B/text_rule.md | {
"path": "tasks/E10/B/real.mp4",
"sha256": "364d415b206fb7234c8d1b778ca9c026eb550dcba5ab49c0e59fe207efbc6ec4",
"bytes": 5702744,
"duration_seconds": 8.466667,
"frame_count": 254,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "37ab59... | {
"path": "tasks/E10/B/sim.mp4",
"sha256": "2eb0dba69f5d0a839b7ab331c05fc884c3cf796027ff8f6f58176cf07d0da1b9",
"bytes": 1378777,
"duration_seconds": 68.8,
"frame_count": 1720,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3e99556dcfac... | available | # E10B · 按属性选择并套柱
## 执行任务目标
按示范的颜色、形状选择方式,把符合条件的零件套到匹配的立柱上,其余留在桌上。
## 原始局部示范任务
按示范的颜色、形状选择方式,把符合条件的零件套到匹配的立柱上,其余留在桌上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E10B 文字规则
套入所有方形件,共3件
|
E10 | C | 按属性选择并套柱 | tasks/E10/C/task.json | tasks/E10/C/prompt.md | tasks/E10/C/text_rule.md | {
"path": "tasks/E10/C/real.mp4",
"sha256": "d548f901316397c7824f2e70b09a7f2f0dc8649424ff6e2f48c7426980967083",
"bytes": 10864706,
"duration_seconds": 15.5,
"frame_count": 465,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "776fe9a12... | {
"path": "tasks/E10/C/sim.mp4",
"sha256": "8ea730390a06d00cece9b689fb93ba07ca58db8d393c0c1eacbf539fcf76bfff",
"bytes": 2664299,
"duration_seconds": 126.36,
"frame_count": 3159,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "80c410b3a0... | available | # E10C · 按属性选择并套柱
## 执行任务目标
按示范的颜色、形状选择方式,把符合条件的零件套到匹配的立柱上,其余留在桌上。
## 原始局部示范任务
按示范的颜色、形状选择方式,把符合条件的零件套到匹配的立柱上,其余留在桌上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E10C 文字规则
套入红色件或方形件,共5件;红方只处理一次
|
E11 | A | 四格布局变换入框 | tasks/E11/A/task.json | tasks/E11/A/prompt.md | tasks/E11/A/text_rule.md | {
"path": "tasks/E11/A/real.mp4",
"sha256": "55a9da0e6c18b38d16856eaad8c3287264efca51331dba81ca5804b83d4133ac",
"bytes": 6820622,
"duration_seconds": 10.433333,
"frame_count": 313,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "17541... | {
"path": "tasks/E11/A/sim.mp4",
"sha256": "a806fa6f15913d44ef5f452746513aab762f6648a88df831509de91b46f82080",
"bytes": 2185113,
"duration_seconds": 85.28,
"frame_count": 2132,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "38eed14d0ee... | available | # E11A · 四格布局变换入框
## 执行任务目标
按示范中的平面布局变换,将开局的四块积木搬入四格框。
## 原始局部示范任务
按示范中的平面布局变换,将开局的四块积木搬入四格框。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E11A 文字规则
保留初始2×2布局搬入四格框
|
E11 | B | 四格布局变换入框 | tasks/E11/B/task.json | tasks/E11/B/prompt.md | tasks/E11/B/text_rule.md | {
"path": "tasks/E11/B/real.mp4",
"sha256": "6d453ae37574d43d6d798b8d3a0505ab213daab2196ee45dd7a5a09c1a40b2a3",
"bytes": 8609603,
"duration_seconds": 12.5,
"frame_count": 375,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "dda6ad81b9... | {
"path": "tasks/E11/B/sim.mp4",
"sha256": "cca1e5e670bb082b0690eb92d79c5a837cf320633fb1024c0e23de8cf7a3fd32",
"bytes": 1837047,
"duration_seconds": 72.52,
"frame_count": 1813,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "809d2d5a6b8... | available | # E11B · 四格布局变换入框
## 执行任务目标
按示范中的平面布局变换,将开局的四块积木搬入四格框。
## 原始局部示范任务
按示范中的平面布局变换,将开局的四块积木搬入四格框。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E11B 文字规则
初始布局顺时针旋转90°后入框
|
E11 | C | 四格布局变换入框 | tasks/E11/C/task.json | tasks/E11/C/prompt.md | tasks/E11/C/text_rule.md | {
"path": "tasks/E11/C/real.mp4",
"sha256": "adfb13eca8e50cab4230ac21b78a2aab3c24a873edcdc4e5aec8fb4d77e73dca",
"bytes": 7196915,
"duration_seconds": 10.6,
"frame_count": 318,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "f9434a9dd8... | {
"path": "tasks/E11/C/sim.mp4",
"sha256": "7f6bc833deaff0781b71ac502a3ce0a4df5f854106148e299d7394e9aa59672b",
"bytes": 2201246,
"duration_seconds": 86.28,
"frame_count": 2157,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "6729f8a0a25... | available | # E11C · 四格布局变换入框
## 执行任务目标
按示范中的平面布局变换,将开局的四块积木搬入四格框。
## 原始局部示范任务
按示范中的平面布局变换,将开局的四块积木搬入四格框。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E11C 文字规则
初始布局左右镜像后入框
|
E12 | A | 插排孔位选择 | tasks/E12/A/task.json | tasks/E12/A/prompt.md | tasks/E12/A/text_rule.md | {
"path": "tasks/E12/A/real.mp4",
"sha256": "f45c73f2aed78f14d784cd6e869c54e5c0341960a14a3f81d11c7143a9965b1c",
"bytes": 4141141,
"duration_seconds": 6.7,
"frame_count": 201,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "4a96ef58ddd... | {
"path": "tasks/E12/A/sim.mp4",
"sha256": "bf44fbb39c347c7fd5c9266f67b55a2a8893cce45a12b4dcacc11c3fe0eff17f",
"bytes": 1872062,
"duration_seconds": 71.8,
"frame_count": 1795,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "b1f9d83724e6... | available | # E12A · 插排孔位选择
## 执行任务目标
按示范相对于电线端的插位选择,将两个充电器插入四位插排。
## 原始局部示范任务
按示范相对于电线端的插位选择,将两个充电器插入四位插排。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E12A 文字规则
两充电器插入靠电线端的两个插位
|
E12 | B | 插排孔位选择 | tasks/E12/B/task.json | tasks/E12/B/prompt.md | tasks/E12/B/text_rule.md | {
"path": "tasks/E12/B/real.mp4",
"sha256": "76cf905f105f2207555236b4672830281cb985d85fd519e0e678e3ef4ed50f69",
"bytes": 4882646,
"duration_seconds": 6.733333,
"frame_count": 202,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "0ed659... | {
"path": "tasks/E12/B/sim.mp4",
"sha256": "8c56450c21d0f8f0c41c33575de6b4af3711edcf715cdc60ea277109c482351b",
"bytes": 1756388,
"duration_seconds": 69.4,
"frame_count": 1735,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9ff9724a1a58... | available | # E12B · 插排孔位选择
## 执行任务目标
按示范相对于电线端的插位选择,将两个充电器插入四位插排。
## 原始局部示范任务
按示范相对于电线端的插位选择,将两个充电器插入四位插排。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E12B 文字规则
插入远离电线端的两个插位
|
E12 | C | 插排孔位选择 | tasks/E12/C/task.json | tasks/E12/C/prompt.md | tasks/E12/C/text_rule.md | {
"path": "tasks/E12/C/real.mp4",
"sha256": "13ef52a770177eb1d310e7ad5941dcc112311ecbf9d2e3d6cb1be36fa1cd54b8",
"bytes": 4239743,
"duration_seconds": 5.933333,
"frame_count": 178,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ba00fd... | {
"path": "tasks/E12/C/sim.mp4",
"sha256": "5f2b53e1ccd7ab8b86c7622ecb9a6ac40334df7b56004400d506b2a4fb1a64e0",
"bytes": 1805019,
"duration_seconds": 69.88,
"frame_count": 1747,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "eca1190cdf2... | available | # E12C · 插排孔位选择
## 执行任务目标
按示范相对于电线端的插位选择,将两个充电器插入四位插排。
## 原始局部示范任务
按示范相对于电线端的插位选择,将两个充电器插入四位插排。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E12C 文字规则
插入两端插位,中间两个空着
|
E13 | A | 手机支架姿态 | tasks/E13/A/task.json | tasks/E13/A/prompt.md | tasks/E13/A/text_rule.md | {
"path": "tasks/E13/A/real.mp4",
"sha256": "86e0cf34a7dd120c29f0932e4a258449b3c5eddb13ce05e70b7dc78fdfebff2c",
"bytes": 3042420,
"duration_seconds": 4.1,
"frame_count": 123,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "a6a89041c49... | {
"path": "tasks/E13/A/sim.mp4",
"sha256": "af7b5502fc0b32d629884d0ae73a1d9a7fc989ea94eed25e77f42ac16dc6f369",
"bytes": 1208126,
"duration_seconds": 52.84,
"frame_count": 1321,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "0acbad906d3... | available | # E13A · 手机支架姿态
## 执行任务目标
按示范的手机朝向,把手机放稳在支架上;以支架开口侧判断左右。
## 原始局部示范任务
按示范的手机朝向,把手机放稳在支架上;以支架开口侧判断左右。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E13A 文字规则
竖放,后摄像头端朝上
|
E13 | B | 手机支架姿态 | tasks/E13/B/task.json | tasks/E13/B/prompt.md | tasks/E13/B/text_rule.md | {
"path": "tasks/E13/B/real.mp4",
"sha256": "b0f7a1f93d349922199e2322c42a8ac923e330853d3617686b692b92ca057f71",
"bytes": 3608206,
"duration_seconds": 5.066667,
"frame_count": 152,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "718a82... | {
"path": "tasks/E13/B/sim.mp4",
"sha256": "94c8a3f311f5f976f9c81c126534231612c2610106cad2e20eaf2746d5385935",
"bytes": 1663772,
"duration_seconds": 71.16,
"frame_count": 1779,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5b5b424c361... | available | # E13B · 手机支架姿态
## 执行任务目标
按示范的手机朝向,把手机放稳在支架上;以支架开口侧判断左右。
## 原始局部示范任务
按示范的手机朝向,把手机放稳在支架上;以支架开口侧判断左右。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E13B 文字规则
横放,后摄像头端朝右
|
E13 | C | 手机支架姿态 | tasks/E13/C/task.json | tasks/E13/C/prompt.md | tasks/E13/C/text_rule.md | {
"path": "tasks/E13/C/real.mp4",
"sha256": "a35b92edbd8e56168a12ed43662bfbe4a3eb04d2c5527483a8d90403b347af25",
"bytes": 3542489,
"duration_seconds": 4.866667,
"frame_count": 146,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "e5619c... | {
"path": "tasks/E13/C/sim.mp4",
"sha256": "24b51279492256740fed6ae8ac7cfacb7b0dfb1c5c298ba9e536e97cc66e5ffa",
"bytes": 1068088,
"duration_seconds": 52.28,
"frame_count": 1307,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "31c30e985ec... | available | # E13C · 手机支架姿态
## 执行任务目标
按示范的手机朝向,把手机放稳在支架上;以支架开口侧判断左右。
## 原始局部示范任务
按示范的手机朝向,把手机放稳在支架上;以支架开口侧判断左右。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E13C 文字规则
横放,后摄像头端朝左
|
E14 | A | 拆塔与层序重排 | tasks/E14/A/task.json | tasks/E14/A/prompt.md | tasks/E14/A/text_rule.md | {
"path": "tasks/E14/A/real.mp4",
"sha256": "92651a7fe89128d958bd9ec8e35493c0e33e4b65a5c4d2411ab8c99409c522d2",
"bytes": 6241889,
"duration_seconds": 9.333333,
"frame_count": 280,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "608915... | {
"path": "tasks/E14/A/sim.mp4",
"sha256": "72f8d19637be60dde0b68d9208af36317d2d09db56f0cc27784d660e9e0f86ee",
"bytes": 2124465,
"duration_seconds": 73.16,
"frame_count": 1829,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "f053f637056... | available | # E14A · 拆塔与层序重排
## 执行任务目标
按示范中对塔层序的调整方式,重新搭好这座四层积木塔。
## 原始局部示范任务
按示范中对塔层序的调整方式,重新搭好这座四层积木塔。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E14A 文字规则
交换从顶端数第1、2层,其余不变
|
E14 | B | 拆塔与层序重排 | tasks/E14/B/task.json | tasks/E14/B/prompt.md | tasks/E14/B/text_rule.md | {
"path": "tasks/E14/B/real.mp4",
"sha256": "40b18eb74ee9d968f0036dd07f270789acd8e9e81c4835e93fa14802f388a10f",
"bytes": 8528471,
"duration_seconds": 12.866667,
"frame_count": 386,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "f8e14... | {
"path": "tasks/E14/B/sim.mp4",
"sha256": "96acd72bcddd9a79d880e158268e7df338f2e87aa6dc89faf44de5a119b9e4fc",
"bytes": 2121704,
"duration_seconds": 73.16,
"frame_count": 1829,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "d685db84b8a... | available | # E14B · 拆塔与层序重排
## 执行任务目标
按示范中对塔层序的调整方式,重新搭好这座四层积木塔。
## 原始局部示范任务
按示范中对塔层序的调整方式,重新搭好这座四层积木塔。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E14B 文字规则
交换从顶端数第2、3层,其余不变
|
E14 | C | 拆塔与层序重排 | tasks/E14/C/task.json | tasks/E14/C/prompt.md | tasks/E14/C/text_rule.md | {
"path": "tasks/E14/C/real.mp4",
"sha256": "5a70a800c9cdc69f27a26a0b0776d81fa7759d52d8e73409815ff8b033472d63",
"bytes": 6949752,
"duration_seconds": 11.166667,
"frame_count": 335,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "bd378... | {
"path": "tasks/E14/C/sim.mp4",
"sha256": "6fbbb97b5977b5af9521707510f134d46524c37719fb20c280aa5b2b1c05ab47",
"bytes": 2181996,
"duration_seconds": 75.2,
"frame_count": 1880,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9252c9768f3d... | available | # E14C · 拆塔与层序重排
## 执行任务目标
按示范中对塔层序的调整方式,重新搭好这座四层积木塔。
## 原始局部示范任务
按示范中对塔层序的调整方式,重新搭好这座四层积木塔。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E14C 文字规则
整塔层序上下反转
|
E15 | A | T形块属性交换 | tasks/E15/A/task.json | tasks/E15/A/prompt.md | tasks/E15/A/text_rule.md | {
"path": "tasks/E15/A/real.mp4",
"sha256": "cae0c834e30477e10b3f677e6e6aad33f5659f9376f9ddc3128155ee8a7fe17a",
"bytes": 7329631,
"duration_seconds": 10.8,
"frame_count": 324,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "307131f3c1... | {
"path": "tasks/E15/A/sim.mp4",
"sha256": "5443a6b153d1e3d7a30106befb77c5af075f5c314fe55a39822e3c4c02d4c125",
"bytes": 1069141,
"duration_seconds": 43.96,
"frame_count": 1099,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "2fd8ab7a913... | available | # E15A · T形块属性交换
## 执行任务目标
按照示范中位置与朝向的交换方式,调整这两个T形块。
## 原始局部示范任务
按照示范中位置与朝向的交换方式,调整这两个T形块。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E15A 文字规则
交换位置,各自朝向不变
|
E15 | B | T形块属性交换 | tasks/E15/B/task.json | tasks/E15/B/prompt.md | tasks/E15/B/text_rule.md | {
"path": "tasks/E15/B/real.mp4",
"sha256": "9e168c6569562d90f5722b04b2c58ed3a4bccde009d86763274fedc89ff19dd2",
"bytes": 4776902,
"duration_seconds": 7.566667,
"frame_count": 227,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c57f16... | {
"path": "tasks/E15/B/sim.mp4",
"sha256": "544255fecbae828be21e33516887e3bf8ef82b591ed1c7b3b8ab597f1346b3a2",
"bytes": 762841,
"duration_seconds": 31.96,
"frame_count": 799,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "e1aee962f6cba... | available | # E15B · T形块属性交换
## 执行任务目标
按照示范中位置与朝向的交换方式,调整这两个T形块。
## 原始局部示范任务
按照示范中位置与朝向的交换方式,调整这两个T形块。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E15B 文字规则
各自位置不变,只交换朝向
|
E15 | C | T形块属性交换 | tasks/E15/C/task.json | tasks/E15/C/prompt.md | tasks/E15/C/text_rule.md | {
"path": "tasks/E15/C/real.mp4",
"sha256": "77712adc93fff599918d3a0acc0542a870d0b1342b6ac30bd528ff1c5219fcde",
"bytes": 11019278,
"duration_seconds": 15.7,
"frame_count": 471,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "35aceda7f... | {
"path": "tasks/E15/C/sim.mp4",
"sha256": "a0b8ea50655dacb0dcbc20303f4732c4ed84029d32a335e48d1e5c48f98b33b9",
"bytes": 1362465,
"duration_seconds": 54.36,
"frame_count": 1359,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "054e1e616a0... | available | # E15C · T形块属性交换
## 执行任务目标
按照示范中位置与朝向的交换方式,调整这两个T形块。
## 原始局部示范任务
按照示范中位置与朝向的交换方式,调整这两个T形块。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E15C 文字规则
位置与朝向都交换
|
E16 | A | 三件布局搬入托盘 | tasks/E16/A/task.json | tasks/E16/A/prompt.md | tasks/E16/A/text_rule.md | {
"path": "tasks/E16/A/real.mp4",
"sha256": "2defa0ea51e863cc385c2da489cdb33aa29e0eb7749dcc01990126008d854002",
"bytes": 8453583,
"duration_seconds": 12.266667,
"frame_count": 368,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "fd89e... | {
"path": "tasks/E16/A/sim.mp4",
"sha256": "a352c5fac4e8ed3731064a9bf67286eebce9fc27c5b753d2386215b1d8a59435",
"bytes": 1361589,
"duration_seconds": 51.08,
"frame_count": 1277,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "f3b8f83d3bc... | available | # E16A · 三件布局搬入托盘
## 执行任务目标
按示范的布局与单件朝向变换,把三件T形块搬进托盘。
## 原始局部示范任务
按示范的布局与单件朝向变换,把三件T形块搬进托盘。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E16A 文字规则
保留三件的相对位置与朝向搬入托盘
|
E16 | B | 三件布局搬入托盘 | tasks/E16/B/task.json | tasks/E16/B/prompt.md | tasks/E16/B/text_rule.md | {
"path": "tasks/E16/B/real.mp4",
"sha256": "81e4ef18602e6b4ada6d73d7e04203008ee140e84176d88350b0f04043992820",
"bytes": 14163651,
"duration_seconds": 19.733333,
"frame_count": 592,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "aee0... | {
"path": "tasks/E16/B/sim.mp4",
"sha256": "d50f12ff87fae540c9c391ee434d5c21f784706bd279bbc6d2a2a5d713ae8dfb",
"bytes": 2672390,
"duration_seconds": 110.44,
"frame_count": 2761,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "a0cdfbec10... | available | # E16B · 三件布局搬入托盘
## 执行任务目标
按示范的布局与单件朝向变换,把三件T形块搬进托盘。
## 原始局部示范任务
按示范的布局与单件朝向变换,把三件T形块搬进托盘。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E16B 文字规则
相对位置不变,每件原地逆时针转90°后搬入
|
E16 | C | 三件布局搬入托盘 | tasks/E16/C/task.json | tasks/E16/C/prompt.md | tasks/E16/C/text_rule.md | {
"path": "tasks/E16/C/real.mp4",
"sha256": "ad8bb734930ed680d73888706cb39e2b72915195e542d41b90cff7c227de2f64",
"bytes": 7555865,
"duration_seconds": 9.966667,
"frame_count": 299,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c2cd55... | {
"path": "tasks/E16/C/sim.mp4",
"sha256": "c985572e61d019c96d7d1da86cc2f9922e5dae6867dbbe16d24beb2ffcf9e9bd",
"bytes": 1398082,
"duration_seconds": 51.84,
"frame_count": 1296,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c172323fc49... | available | # E16C · 三件布局搬入托盘
## 执行任务目标
按示范的布局与单件朝向变换,把三件T形块搬进托盘。
## 原始局部示范任务
按示范的布局与单件朝向变换,把三件T形块搬进托盘。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E16C 文字规则
相对位置上下镜像、各件朝向不变后搬入
|
E17 | A | 展示架补货与开场准备 | tasks/E17/A/task.json | tasks/E17/A/prompt.md | tasks/E17/A/text_rule.md | {
"path": "tasks/E17/A/real.mp4",
"sha256": "f434e20b8c2835429f2e996083623d325c41300a07e9af7b3d4810c41d4113dd",
"bytes": 11329458,
"duration_seconds": 15.966667,
"frame_count": 479,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ad60... | {
"path": "tasks/E17/A/sim.mp4",
"sha256": "32938a5f3ff3fd7257f6e8d13e14d73416a82926c1ace1a71b68380ee336b685",
"bytes": 3576153,
"duration_seconds": 127.8,
"frame_count": 3195,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "513eaa20a6a... | available | # E17A · 展示架补货与开场准备
## 执行任务目标
为指定展台完成一次开场备货。按视频的分层对应关系摆放本批展品;展示架必须在订单指定位置,包装收回回收盒,架上不得混入其他批次展品。
## 原始局部示范任务
按照示范的分层收纳方式,把四件物体放到点心架上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E17A 文字规则
蓝色放上层,绿色放下层
|
E17 | B | 展示架补货与开场准备 | tasks/E17/B/task.json | tasks/E17/B/prompt.md | tasks/E17/B/text_rule.md | {
"path": "tasks/E17/B/real.mp4",
"sha256": "cdb218011906dd31b3c75080213798c52a32f92b85ff16a670c4d961f1c2ad0a",
"bytes": 11872623,
"duration_seconds": 16.4,
"frame_count": 492,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "88f9d3803... | {
"path": "tasks/E17/B/sim.mp4",
"sha256": "d58be0776affc33feadb8643fc7ee10ef5f1fa4f7f6be56c818fb2efe2c97078",
"bytes": 4010982,
"duration_seconds": 149.28,
"frame_count": 3732,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "227f0c88fa... | available | # E17B · 展示架补货与开场准备
## 执行任务目标
为指定展台完成一次开场备货。按视频的分层对应关系摆放本批展品;展示架必须在订单指定位置,包装收回回收盒,架上不得混入其他批次展品。
## 原始局部示范任务
按照示范的分层收纳方式,把四件物体放到点心架上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E17B 文字规则
三角放上层,方形放下层
|
E17 | C | 展示架补货与开场准备 | tasks/E17/C/task.json | tasks/E17/C/prompt.md | tasks/E17/C/text_rule.md | {
"path": "tasks/E17/C/real.mp4",
"sha256": "416e80183678ab16561c59d509b3c929f9c490ced113b90cabca528d680279ab",
"bytes": 10278573,
"duration_seconds": 14.7,
"frame_count": 441,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5e50ef869... | {
"path": "tasks/E17/C/sim.mp4",
"sha256": "858dfdc3c4b1a952b64b9334c055847eb1f93876b700b6efbac100f127554dfa",
"bytes": 2929435,
"duration_seconds": 124.56,
"frame_count": 3114,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "d85a9d5cb1... | available | # E17C · 展示架补货与开场准备
## 执行任务目标
为指定展台完成一次开场备货。按视频的分层对应关系摆放本批展品;展示架必须在订单指定位置,包装收回回收盒,架上不得混入其他批次展品。
## 原始局部示范任务
按照示范的分层收纳方式,把四件物体放到点心架上。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E17C 文字规则
四件全部放上层
|
E18 | A | 抽屉取球并关回 | tasks/E18/A/task.json | tasks/E18/A/prompt.md | tasks/E18/A/text_rule.md | {
"path": "tasks/E18/A/real.mp4",
"sha256": "9369535410f30c888c70726a6b890e99dc663cebe3ccd0288c5f5b53f6d11eb8",
"bytes": 9230670,
"duration_seconds": 12.133333,
"frame_count": 364,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "39b1d... | {
"path": "tasks/E18/A/sim.mp4",
"sha256": "f5fe4e7eb70abb495cc4138a31cbbbe6057e4269817ca03fd3e564ddef1d5b8a",
"bytes": 943600,
"duration_seconds": 36.44,
"frame_count": 911,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c5c74e73da96f... | available | # E18A · 抽屉取球并关回
## 执行任务目标
按示范选择同色的球,取出放到白色垫上,并关好抽屉。
## 原始局部示范任务
按示范选择同色的球,取出放到白色垫上,并关好抽屉。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E18A 文字规则
取出红球放到白色目标纸上,关回抽屉
|
E18 | B | 抽屉取球并关回 | tasks/E18/B/task.json | tasks/E18/B/prompt.md | tasks/E18/B/text_rule.md | {
"path": "tasks/E18/B/real.mp4",
"sha256": "9c4ad9cfc197e002695045f4f23306cd277218cbc5ba01ae546050ac6ac05056",
"bytes": 9708620,
"duration_seconds": 12.8,
"frame_count": 384,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c6818f3115... | {
"path": "tasks/E18/B/sim.mp4",
"sha256": "b6c5c37e81eab36fda12219e9ca9565740d57d5f4f7975db0dcfa908744c03a0",
"bytes": 1070444,
"duration_seconds": 45.36,
"frame_count": 1134,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "fd0da5c29ea... | available | # E18B · 抽屉取球并关回
## 执行任务目标
按示范选择同色的球,取出放到白色垫上,并关好抽屉。
## 原始局部示范任务
按示范选择同色的球,取出放到白色垫上,并关好抽屉。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E18B 文字规则
取出绿球放到白色目标纸上,关回抽屉
|
E18 | C | 抽屉取球并关回 | tasks/E18/C/task.json | tasks/E18/C/prompt.md | tasks/E18/C/text_rule.md | {
"path": "tasks/E18/C/real.mp4",
"sha256": "6b01c4498d45355055030ff9333a282cdf7523723d78e10c58fcba0d617696f1",
"bytes": 8347664,
"duration_seconds": 10.966667,
"frame_count": 329,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9374f... | {
"path": "tasks/E18/C/sim.mp4",
"sha256": "cc3223d8437cf55b85e634e318444651c839bbafd9fea4ef05222939ee8bfdb4",
"bytes": 1084643,
"duration_seconds": 45.08,
"frame_count": 1127,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "764f2bceda2... | available | # E18C · 抽屉取球并关回
## 执行任务目标
按示范选择同色的球,取出放到白色垫上,并关好抽屉。
## 原始局部示范任务
按示范选择同色的球,取出放到白色垫上,并关好抽屉。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E18C 文字规则
取出蓝球放到白色目标纸上,关回抽屉
|
E19 | A | 饮料瓶与瓶盖配对 | tasks/E19/A/task.json | tasks/E19/A/prompt.md | tasks/E19/A/text_rule.md | {
"path": "tasks/E19/A/real.mp4",
"sha256": "3ca16c19544c640f733160bfa300544657cf37f795612b2893894857c87626d9",
"bytes": 11837019,
"duration_seconds": 16.2,
"frame_count": 486,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "65714a4d4... | {
"path": "tasks/E19/A/sim.mp4",
"sha256": "72fa6c94bca5382b74e39ae9eb23f5800717192745482178edbf0aa4adb98b91",
"bytes": 1255409,
"duration_seconds": 53.4,
"frame_count": 1335,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5e0cb49332ec... | available | # E19A · 饮料瓶与瓶盖配对
## 执行任务目标
按示范中瓶型与瓶盖颜色的对应关系,为三瓶饮料配好盖。
## 原始局部示范任务
按示范中瓶型与瓶盖颜色的对应关系,为三瓶饮料配好盖。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E19A 文字规则
雪碧配黄盖,水配白盖,可乐配红盖
|
E19 | B | 饮料瓶与瓶盖配对 | tasks/E19/B/task.json | tasks/E19/B/prompt.md | tasks/E19/B/text_rule.md | {
"path": "tasks/E19/B/real.mp4",
"sha256": "d99ab3cef29062abd41291bcf7f27316f6879326debe986dc88f26e10c39f50c",
"bytes": 10787693,
"duration_seconds": 15.266667,
"frame_count": 458,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "797d... | {
"path": "tasks/E19/B/sim.mp4",
"sha256": "a2c35b938368866f21b2f61f89b231d920259275bf263173750ccaee83274f1c",
"bytes": 1530380,
"duration_seconds": 57.96,
"frame_count": 1449,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3dfc7cbeb5f... | available | # E19B · 饮料瓶与瓶盖配对
## 执行任务目标
按示范中瓶型与瓶盖颜色的对应关系,为三瓶饮料配好盖。
## 原始局部示范任务
按示范中瓶型与瓶盖颜色的对应关系,为三瓶饮料配好盖。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E19B 文字规则
雪碧配红盖,水配黄盖,可乐配白盖
|
E19 | C | 饮料瓶与瓶盖配对 | tasks/E19/C/task.json | tasks/E19/C/prompt.md | tasks/E19/C/text_rule.md | {
"path": "tasks/E19/C/real.mp4",
"sha256": "5811379d44a9da96cf8cc6119bf9d6f3386f65c351021fb2a0b33f118120d034",
"bytes": 10297755,
"duration_seconds": 13.933333,
"frame_count": 418,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5412... | {
"path": "tasks/E19/C/sim.mp4",
"sha256": "3a4cc6ee7d18709a7778a4835ad05db919b1b1698261783b1769f7adc18daebc",
"bytes": 1605552,
"duration_seconds": 67.48,
"frame_count": 1687,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "07cd8c2917b... | available | # E19C · 饮料瓶与瓶盖配对
## 执行任务目标
按示范中瓶型与瓶盖颜色的对应关系,为三瓶饮料配好盖。
## 原始局部示范任务
按示范中瓶型与瓶盖颜色的对应关系,为三瓶饮料配好盖。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E19C 文字规则
雪碧配白盖,水配红盖,可乐配黄盖
|
E20 | A | 耗材盘补量后封装交付 | tasks/E20/A/task.json | tasks/E20/A/prompt.md | tasks/E20/A/text_rule.md | {
"path": "tasks/E20/A/real.mp4",
"sha256": "3cb440d7e5a2184bd3f1dcb926aac6291eb815b30153453eaa333b639f128971",
"bytes": 10309619,
"duration_seconds": 14.566667,
"frame_count": 437,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5cc1... | {
"path": "tasks/E20/A/sim.mp4",
"sha256": "50839766538fb3748e16053382ddda147bf95bcbc04b6007d6c1cf38054bffbe",
"bytes": 1801668,
"duration_seconds": 66.6,
"frame_count": 1665,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "63bad7c7b65f... | available | # E20A · 耗材盘补量后封装交付
## 执行任务目标
把订单指定的耗材盘准备成可直接使用的一套并送往指定工位。各图形区的配色和补齐后的总数量按视频;已有合格耗材计入总数,不得清空重装,备用耗材不应随盘送出。
## 原始局部示范任务
按照示范中的配色与补齐后的数量,将各标记区域的筹码补齐。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E20A 文字规则
圆标记黄片、方标记红片、三角标记蓝片,各补到2片
|
E20 | B | 耗材盘补量后封装交付 | tasks/E20/B/task.json | tasks/E20/B/prompt.md | tasks/E20/B/text_rule.md | {
"path": "tasks/E20/B/real.mp4",
"sha256": "0317da91ef314079dfe6c1fe4e811bde1bf4adb96731649ef1d28515017914f2",
"bytes": 16054634,
"duration_seconds": 22.4,
"frame_count": 672,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c7231f2a1... | {
"path": "tasks/E20/B/sim.mp4",
"sha256": "36ec2cbc58eab5548d195bf4aede52eac61bc012b9cdd06fef80cbb0a2a42c97",
"bytes": 2958249,
"duration_seconds": 105.08,
"frame_count": 2627,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "13099b53bf... | available | # E20B · 耗材盘补量后封装交付
## 执行任务目标
把订单指定的耗材盘准备成可直接使用的一套并送往指定工位。各图形区的配色和补齐后的总数量按视频;已有合格耗材计入总数,不得清空重装,备用耗材不应随盘送出。
## 原始局部示范任务
按照示范中的配色与补齐后的数量,将各标记区域的筹码补齐。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E20B 文字规则
同样配色,各补到3片
|
E20 | C | 耗材盘补量后封装交付 | tasks/E20/C/task.json | tasks/E20/C/prompt.md | tasks/E20/C/text_rule.md | {
"path": "tasks/E20/C/real.mp4",
"sha256": "4d0e402e6769bac8f22dff0025c4940d9ec5a7880b4abbe2abcd80506480e3c0",
"bytes": 21435791,
"duration_seconds": 29.566667,
"frame_count": 887,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "55ef... | {
"path": "tasks/E20/C/sim.mp4",
"sha256": "66b1e528e64cf34ed3561d6bb9688d68b033a3cd9cc39956111739247b923855",
"bytes": 4082835,
"duration_seconds": 147.08,
"frame_count": 3677,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "8f0784f9c0... | available | # E20C · 耗材盘补量后封装交付
## 执行任务目标
把订单指定的耗材盘准备成可直接使用的一套并送往指定工位。各图形区的配色和补齐后的总数量按视频;已有合格耗材计入总数,不得清空重装,备用耗材不应随盘送出。
## 原始局部示范任务
按照示范中的配色与补齐后的数量,将各标记区域的筹码补齐。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E20C 文字规则
同样配色,各补到4片
|
E21 | A | 按颜色顺序装四格盘 | tasks/E21/A/task.json | tasks/E21/A/prompt.md | tasks/E21/A/text_rule.md | {
"path": "tasks/E21/A/real.mp4",
"sha256": "89bacc06dc5848f4bb84ef52c052f47f72ee6abeeda681c73c68b106d367e8a9",
"bytes": 5905219,
"duration_seconds": 8.1,
"frame_count": 243,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "68751d98ce4... | {
"path": "tasks/E21/A/sim.mp4",
"sha256": "5f41286ad34672d8ef8c4027edb0fa93625fbfea39cca144cb35e8f337fab836",
"bytes": 2211085,
"duration_seconds": 83.16,
"frame_count": 2079,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "f9674d5e9f8... | available | # E21A · 按颜色顺序装四格盘
## 执行任务目标
按示范的颜色先后顺序,依次将四块积木装进四格盘。
## 原始局部示范任务
按示范的颜色先后顺序,依次将四块积木装进四格盘。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E21A 文字规则
按红、黄、蓝、绿顺序放入左上、右上、左下、右下
|
E21 | B | 按颜色顺序装四格盘 | tasks/E21/B/task.json | tasks/E21/B/prompt.md | tasks/E21/B/text_rule.md | {
"path": "tasks/E21/B/real.mp4",
"sha256": "18a423dc24833ad90a70d0d73f8c588cfaf1f6d65315faff183ebff9c44df29f",
"bytes": 6240226,
"duration_seconds": 8.3,
"frame_count": 249,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "e67bd1b7d0b... | {
"path": "tasks/E21/B/sim.mp4",
"sha256": "64a7e6bf927a4a9fa1ce4b462579d1184a6d5d157117da3c6b8eb25854eae8e3",
"bytes": 2046704,
"duration_seconds": 76.8,
"frame_count": 1920,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "a2543fd1dbd7... | available | # E21B · 按颜色顺序装四格盘
## 执行任务目标
按示范的颜色先后顺序,依次将四块积木装进四格盘。
## 原始局部示范任务
按示范的颜色先后顺序,依次将四块积木装进四格盘。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E21B 文字规则
按红、绿、蓝、黄顺序放入这四格
|
E21 | C | 按颜色顺序装四格盘 | tasks/E21/C/task.json | tasks/E21/C/prompt.md | tasks/E21/C/text_rule.md | {
"path": "tasks/E21/C/real.mp4",
"sha256": "4cfdd14ac80a95955adb9d6493d8614e2628379a1a659effe9b11a2fc48e1be1",
"bytes": 7206738,
"duration_seconds": 9.6,
"frame_count": 288,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "fe00adfef56... | {
"path": "tasks/E21/C/sim.mp4",
"sha256": "c8c0def0f29698f9c069b1020f5050f65f7b0c277b5cc81baf7989f2de351492",
"bytes": 1922624,
"duration_seconds": 70.24,
"frame_count": 1756,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "78f8d1645cb... | available | # E21C · 按颜色顺序装四格盘
## 执行任务目标
按示范的颜色先后顺序,依次将四块积木装进四格盘。
## 原始局部示范任务
按示范的颜色先后顺序,依次将四块积木装进四格盘。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E21C 文字规则
按绿、蓝、黄、红顺序放入这四格
|
E22 | A | 围绕玩具车摆放色块 | tasks/E22/A/task.json | tasks/E22/A/prompt.md | tasks/E22/A/text_rule.md | {
"path": "tasks/E22/A/real.mp4",
"sha256": "c077cc7ed86f1fdcd3bf062d51b6d244b9ca5774a675b76ab3474182d28d5141",
"bytes": 5113787,
"duration_seconds": 7.766667,
"frame_count": 233,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "88e740... | {
"path": "tasks/E22/A/sim.mp4",
"sha256": "c6e1d5a177466242ca85ce13868bc14046b50e55ec9240ac56b84500ef1a785a",
"bytes": 1593972,
"duration_seconds": 60.64,
"frame_count": 1516,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "e78c1ff77e9... | available | # E22A · 围绕玩具车摆放色块
## 执行任务目标
按示范中颜色与车前、车左、车右的对应关系,将三块积木摆到玩具车周围。
## 原始局部示范任务
按示范中颜色与车前、车左、车右的对应关系,将三块积木摆到玩具车周围。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E22A 文字规则
车前红、车左绿、车右蓝
|
E22 | B | 围绕玩具车摆放色块 | tasks/E22/B/task.json | tasks/E22/B/prompt.md | tasks/E22/B/text_rule.md | {
"path": "tasks/E22/B/real.mp4",
"sha256": "80b0b88c46227e4a8c2894ddf3a9050371578d933c32a31ac635663d7b0fac53",
"bytes": 6835554,
"duration_seconds": 10.066667,
"frame_count": 302,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "6ff55... | {
"path": "tasks/E22/B/sim.mp4",
"sha256": "fcce0243fce6147f8af1523283f2cd76d872a2c57a743a22086fceab951eef4a",
"bytes": 1500371,
"duration_seconds": 58.72,
"frame_count": 1468,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "0b473a49563... | available | # E22B · 围绕玩具车摆放色块
## 执行任务目标
按示范中颜色与车前、车左、车右的对应关系,将三块积木摆到玩具车周围。
## 原始局部示范任务
按示范中颜色与车前、车左、车右的对应关系,将三块积木摆到玩具车周围。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E22B 文字规则
车前绿、车左蓝、车右红
|
E22 | C | 围绕玩具车摆放色块 | tasks/E22/C/task.json | tasks/E22/C/prompt.md | tasks/E22/C/text_rule.md | {
"path": "tasks/E22/C/real.mp4",
"sha256": "21a69fea456551d003359275ad1b32cbed0303486b40840f87df0f47ba3a9d7f",
"bytes": 4765667,
"duration_seconds": 7.3,
"frame_count": 219,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "46ff035d920... | {
"path": "tasks/E22/C/sim.mp4",
"sha256": "b153373e8c71c0f6aa2971f4aeae307a0464c9c7c0e3e6a55395e95582b96da7",
"bytes": 1507756,
"duration_seconds": 58.68,
"frame_count": 1467,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "4e5de8d5e5b... | available | # E22C · 围绕玩具车摆放色块
## 执行任务目标
按示范中颜色与车前、车左、车右的对应关系,将三块积木摆到玩具车周围。
## 原始局部示范任务
按示范中颜色与车前、车左、车右的对应关系,将三块积木摆到玩具车周围。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E22C 文字规则
车前蓝、车左红、车右绿
|
E23 | A | 按层位将彩杯套上杯架 | tasks/E23/A/task.json | tasks/E23/A/prompt.md | tasks/E23/A/text_rule.md | {
"path": "tasks/E23/A/real.mp4",
"sha256": "0eaa40a2a02eb956d78d4460d313d9d516007c73c5ad68729dc53143245fb8bb",
"bytes": 6098510,
"duration_seconds": 8.833333,
"frame_count": 265,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3865d6... | {
"path": "tasks/E23/A/sim.mp4",
"sha256": "dddd017080142d886465242f368e93c77450bca0276dd7f50de1510bf5f962c2",
"bytes": 6194012,
"duration_seconds": 240.12,
"frame_count": 6003,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "6007f2b46e... | available | # E23A · 按层位将彩杯套上杯架
## 执行任务目标
按示范的颜色高低次序,把三个杯子套到杯架的三个高度层。
## 原始局部示范任务
按示范的颜色高低次序,把三个杯子套到杯架的三个高度层。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E23A 文字规则
从高到低红、绿、蓝
|
E23 | B | 按层位将彩杯套上杯架 | tasks/E23/B/task.json | tasks/E23/B/prompt.md | tasks/E23/B/text_rule.md | {
"path": "tasks/E23/B/real.mp4",
"sha256": "657941c4ff30433f484a68decc7d1c9742c711e5ddbdc2df4b8623285089ba64",
"bytes": 6285234,
"duration_seconds": 8.3,
"frame_count": 249,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "b0d7bcd8c8c... | {
"path": "tasks/E23/B/sim.mp4",
"sha256": "b4285f0edba89be628dc393a056184e64ec1038e0fe7930d9720d462a78d7bb5",
"bytes": 5747047,
"duration_seconds": 225.36,
"frame_count": 5634,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3af5c09e7c... | available | # E23B · 按层位将彩杯套上杯架
## 执行任务目标
按示范的颜色高低次序,把三个杯子套到杯架的三个高度层。
## 原始局部示范任务
按示范的颜色高低次序,把三个杯子套到杯架的三个高度层。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E23B 文字规则
从高到低红、蓝、绿
|
E23 | C | 按层位将彩杯套上杯架 | tasks/E23/C/task.json | tasks/E23/C/prompt.md | tasks/E23/C/text_rule.md | {
"path": "tasks/E23/C/real.mp4",
"sha256": "b66dda8c4d822b7c65dbca3bb2f0480ee8da9bf68a6e6e550816737c0e7ee737",
"bytes": 5941863,
"duration_seconds": 7.933333,
"frame_count": 238,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "590dc7... | {
"path": "tasks/E23/C/sim.mp4",
"sha256": "929c27ee19a7e3d86576a2c74af2ddfdc0ff20ea6b7a8ca1c6cdaf458844d51a",
"bytes": 6620632,
"duration_seconds": 267.84,
"frame_count": 6696,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "fc6dca6ff3... | available | # E23C · 按层位将彩杯套上杯架
## 执行任务目标
按示范的颜色高低次序,把三个杯子套到杯架的三个高度层。
## 原始局部示范任务
按示范的颜色高低次序,把三个杯子套到杯架的三个高度层。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E23C 文字规则
从高到低蓝、绿、红
|
E24 | A | 彩杯按序装盒并盖盖 | tasks/E24/A/task.json | tasks/E24/A/prompt.md | tasks/E24/A/text_rule.md | {
"path": "tasks/E24/A/real.mp4",
"sha256": "913ce8b8024f405655fb2574884121125cea5095188d4743371fd6facdec6154",
"bytes": 10261363,
"duration_seconds": 13.466667,
"frame_count": 404,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "2613... | {
"path": "tasks/E24/A/sim.mp4",
"sha256": "2e0be2da610059364fb1296c5412d896b4335732bf02d8268692634cdc0b74ad",
"bytes": 4393310,
"duration_seconds": 202.92,
"frame_count": 5073,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "174664c2ac... | available | # E24A · 彩杯按序装盒并盖盖
## 执行任务目标
按示范的颜色顺序把三个杯子装入盒中,最后盖好盒盖。
## 原始局部示范任务
按示范的颜色顺序把三个杯子装入盒中,最后盖好盒盖。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E24A 文字规则
红→绿→蓝依次装盒,盖上盒盖
|
E24 | B | 彩杯按序装盒并盖盖 | tasks/E24/B/task.json | tasks/E24/B/prompt.md | tasks/E24/B/text_rule.md | {
"path": "tasks/E24/B/real.mp4",
"sha256": "ef1a2ab3564f50e564611b30dde94ecc87158d595a263ebbf82426b44c8c9ec4",
"bytes": 8323132,
"duration_seconds": 10.833333,
"frame_count": 325,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "81ea8... | {
"path": "tasks/E24/B/sim.mp4",
"sha256": "f95df398b04305dbfc5f5b78a93a5eb72c80337fd3ed9d33b0aae6d91107de5b",
"bytes": 4410961,
"duration_seconds": 203.04,
"frame_count": 5076,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "735783b8b9... | available | # E24B · 彩杯按序装盒并盖盖
## 执行任务目标
按示范的颜色顺序把三个杯子装入盒中,最后盖好盒盖。
## 原始局部示范任务
按示范的颜色顺序把三个杯子装入盒中,最后盖好盒盖。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E24B 文字规则
绿→蓝→红依次装盒,盖上盒盖
|
E24 | C | 彩杯按序装盒并盖盖 | tasks/E24/C/task.json | tasks/E24/C/prompt.md | tasks/E24/C/text_rule.md | {
"path": "tasks/E24/C/real.mp4",
"sha256": "df66bddf39308befa43dd61e1e8c752e553e69997a0edd6fcd2b21739a5b1137",
"bytes": 7780417,
"duration_seconds": 10.2,
"frame_count": 306,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "04328a28c5... | {
"path": "tasks/E24/C/sim.mp4",
"sha256": "26981eb9eeb2cca76c83a458a6d26982f02f5b21c6a8766afc76f9824794bc5f",
"bytes": 4402366,
"duration_seconds": 203,
"frame_count": 5075,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "b0feb4e2a4811... | available | # E24C · 彩杯按序装盒并盖盖
## 执行任务目标
按示范的颜色顺序把三个杯子装入盒中,最后盖好盒盖。
## 原始局部示范任务
按示范的颜色顺序把三个杯子装入盒中,最后盖好盒盖。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E24C 文字规则
蓝→红→绿依次装盒,盖上盒盖
|
E25 | A | 餐具朝向与盘边布局 | tasks/E25/A/task.json | tasks/E25/A/prompt.md | tasks/E25/A/text_rule.md | {
"path": "tasks/E25/A/real.mp4",
"sha256": "aa8e55c6e6d0d45fac9dfcf4dc59e2fb09bdc5cdf389467ab7e9889a7009e8e7",
"bytes": 8755118,
"duration_seconds": 11.666667,
"frame_count": 350,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "cc3e0... | {
"path": "tasks/E25/A/sim.mp4",
"sha256": "6ba888f1e62cd81fccd7be5b745a10738e9af6165b189b11cec399d44b912b9a",
"bytes": 1754069,
"duration_seconds": 70.36,
"frame_count": 1759,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ce9e0ac6fae... | available | # E25A · 餐具朝向与盘边布局
## 执行任务目标
按示范中餐具相对餐盘的位置和端点朝向,布置勺、刀、叉。
## 原始局部示范任务
按示范中餐具相对餐盘的位置和端点朝向,布置勺、刀、叉。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E25A 文字规则
勺、刀、叉排在盘远侧,柄端靠盘
|
E25 | B | 餐具朝向与盘边布局 | tasks/E25/B/task.json | tasks/E25/B/prompt.md | tasks/E25/B/text_rule.md | {
"path": "tasks/E25/B/real.mp4",
"sha256": "41dc7ba669d6a6e6e6302dd223812aeff029a2038f0b31e58b560e4c0b64f4e6",
"bytes": 9033893,
"duration_seconds": 12.033333,
"frame_count": 361,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "90c7a... | {
"path": "tasks/E25/B/sim.mp4",
"sha256": "5687a3c8c2673c80b11a08f7f18b6f4e3ed9da51ac96184361e7e746ea160478",
"bytes": 2386578,
"duration_seconds": 114.96,
"frame_count": 2874,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "bc34eb0c72... | available | # E25B · 餐具朝向与盘边布局
## 执行任务目标
按示范中餐具相对餐盘的位置和端点朝向,布置勺、刀、叉。
## 原始局部示范任务
按示范中餐具相对餐盘的位置和端点朝向,布置勺、刀、叉。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E25B 文字规则
勺、刀、叉排在盘远侧,工作端靠盘
|
E25 | C | 餐具朝向与盘边布局 | tasks/E25/C/task.json | tasks/E25/C/prompt.md | tasks/E25/C/text_rule.md | {
"path": "tasks/E25/C/real.mp4",
"sha256": "84e32ec5d5cfb34403315cd96d2b530d22ec8ac5f60f52edcb52ffd2db1a2e70",
"bytes": 8227405,
"duration_seconds": 10.833333,
"frame_count": 325,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "17777... | {
"path": "tasks/E25/C/sim.mp4",
"sha256": "e14fa19a69df971e16f284730d98ec4e12dbbcf87a23134b97388a3d1a7a2bb9",
"bytes": 1674165,
"duration_seconds": 64.56,
"frame_count": 1614,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "45e46e180d5... | available | # E25C · 餐具朝向与盘边布局
## 执行任务目标
按示范中餐具相对餐盘的位置和端点朝向,布置勺、刀、叉。
## 原始局部示范任务
按示范中餐具相对餐盘的位置和端点朝向,布置勺、刀、叉。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E25C 文字规则
刀放盘左侧,勺和叉放盘右侧,柄朝操作者
|
E26 | A | 数字牌与按铃次数 | tasks/E26/A/task.json | tasks/E26/A/prompt.md | tasks/E26/A/text_rule.md | {
"path": "tasks/E26/A/real.mp4",
"sha256": "d60396b4ad7e7a3e9d0fa159c55757e4cef58d74b074084ba07ceb5bdda617e9",
"bytes": 9418273,
"duration_seconds": 13.566667,
"frame_count": 407,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "5f146... | {
"path": "tasks/E26/A/sim.mp4",
"sha256": "4cb6f93b95e9a4fcb239b98acd88c146c55d3086543b3db8132a7a6e0d6739d8",
"bytes": 523118,
"duration_seconds": 22.72,
"frame_count": 568,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "b269427039a59... | available | # E26A · 数字牌与按铃次数
## 执行任务目标
根据示范展示的数字牌与左右按铃次数之间的对应方式,完成按铃。
## 原始局部示范任务
根据示范展示的数字牌与左右按铃次数之间的对应方式,完成按铃。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E26A 文字规则
左铃按左牌数字,右铃按右牌数字(本片9、1次)
|
E26 | B | 数字牌与按铃次数 | tasks/E26/B/task.json | tasks/E26/B/prompt.md | tasks/E26/B/text_rule.md | {
"path": "tasks/E26/B/real.mp4",
"sha256": "556bf7f702b60aa57d18dd8b9e7c6f0858a7f6fb8bdbc333a2043d17b29918c7",
"bytes": 8360296,
"duration_seconds": 13.133333,
"frame_count": 394,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "4d17a... | {
"path": "tasks/E26/B/sim.mp4",
"sha256": "9e373cca150bd8b9a570488c088bf4ea3686b9de3f0d69ee529b8b2e12aaaf6e",
"bytes": 541874,
"duration_seconds": 22.72,
"frame_count": 568,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "1e551cb427847... | available | # E26B · 数字牌与按铃次数
## 执行任务目标
根据示范展示的数字牌与左右按铃次数之间的对应方式,完成按铃。
## 原始局部示范任务
根据示范展示的数字牌与左右按铃次数之间的对应方式,完成按铃。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E26B 文字规则
左铃按右牌数字,右铃按左牌数字(本片1、9次)
|
E26 | C | 数字牌与按铃次数 | tasks/E26/C/task.json | tasks/E26/C/prompt.md | tasks/E26/C/text_rule.md | {
"path": "tasks/E26/C/real.mp4",
"sha256": "3751dc9a4c8183154746ae245d3b0844c9fb6ca3c650545d5ad5c7f84d218afb",
"bytes": 8607017,
"duration_seconds": 12.466667,
"frame_count": 374,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9d665... | {
"path": "tasks/E26/C/sim.mp4",
"sha256": "396b5870d95edbbdeac605c9b27eedc11989439d9c8389878549144a051ed8d2",
"bytes": 421338,
"duration_seconds": 20.52,
"frame_count": 513,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "b04779e41c6f1... | available | # E26C · 数字牌与按铃次数
## 执行任务目标
根据示范展示的数字牌与左右按铃次数之间的对应方式,完成按铃。
## 原始局部示范任务
根据示范展示的数字牌与左右按铃次数之间的对应方式,完成按铃。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E26C 文字规则
仅左铃按两牌数字之和,右铃不按(本片10、0次)
|
E27 | A | 盘旁勺叉摆放 | tasks/E27/A/task.json | tasks/E27/A/prompt.md | tasks/E27/A/text_rule.md | {
"path": "tasks/E27/A/real.mp4",
"sha256": "9bfe699d2b9a44ff282f30b61475cb810dbdb57205af029f6aadecc43fa89ea5",
"bytes": 4088094,
"duration_seconds": 6.433333,
"frame_count": 193,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "3aa76a... | {
"path": "tasks/E27/A/sim.mp4",
"sha256": "bfa97246fea5a0ac58bf65d2577ad0f0a25cd0ce1ee60a2358a9985686594832",
"bytes": 1423047,
"duration_seconds": 54.8,
"frame_count": 1370,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "40a696e76082... | available | # E27A · 盘旁勺叉摆放
## 执行任务目标
按示范中勺叉相对于餐盘的位置关系摆放餐具。左右以你的用餐侧为准。
## 原始局部示范任务
按示范中勺叉相对于餐盘的位置关系摆放餐具。左右以你的用餐侧为准。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E27A 文字规则
叉在盘左、勺在盘右
|
E27 | B | 盘旁勺叉摆放 | tasks/E27/B/task.json | tasks/E27/B/prompt.md | tasks/E27/B/text_rule.md | {
"path": "tasks/E27/B/real.mp4",
"sha256": "f1316eca773a16883ee4bb6b84b75f10ee80fb4f45bf4efda72b41f1a41efd87",
"bytes": 4907502,
"duration_seconds": 7.9,
"frame_count": 237,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "45f58ca47ec... | {
"path": "tasks/E27/B/sim.mp4",
"sha256": "8f05e352828fa0ce05110a55744604217e6ccc1624d1a90c37168fd2ecf59d0b",
"bytes": 1535629,
"duration_seconds": 58.04,
"frame_count": 1451,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "ea171fb3635... | available | # E27B · 盘旁勺叉摆放
## 执行任务目标
按示范中勺叉相对于餐盘的位置关系摆放餐具。左右以你的用餐侧为准。
## 原始局部示范任务
按示范中勺叉相对于餐盘的位置关系摆放餐具。左右以你的用餐侧为准。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E27B 文字规则
勺在盘左、叉在盘右
|
E27 | C | 盘旁勺叉摆放 | tasks/E27/C/task.json | tasks/E27/C/prompt.md | tasks/E27/C/text_rule.md | {
"path": "tasks/E27/C/real.mp4",
"sha256": "ff460bd97c5c444511504790d3d94a5553343d1b96040ce0d8902cde44602ff7",
"bytes": 4097580,
"duration_seconds": 6.4,
"frame_count": 192,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "16b0d64f650... | {
"path": "tasks/E27/C/sim.mp4",
"sha256": "20bf63025ccd8878ac4e9af6f2d658d4e692d5a86c97c01a28c3f92bace54ed2",
"bytes": 1550624,
"duration_seconds": 58.28,
"frame_count": 1457,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "aa2a2949284... | available | # E27C · 盘旁勺叉摆放
## 执行任务目标
按示范中勺叉相对于餐盘的位置关系摆放餐具。左右以你的用餐侧为准。
## 原始局部示范任务
按示范中勺叉相对于餐盘的位置关系摆放餐具。左右以你的用餐侧为准。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E27C 文字规则
勺叉都放盘左侧,靠盘为勺,外侧为叉
|
E28 | A | 用平铲选取食物装盘 | tasks/E28/A/task.json | tasks/E28/A/prompt.md | tasks/E28/A/text_rule.md | {
"path": "tasks/E28/A/real.mp4",
"sha256": "26345d3e9b6392d25aff7edfdfc8121deb99dcdc94f6901495165e94bfeba429",
"bytes": 3198979,
"duration_seconds": 5.333333,
"frame_count": 160,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "e3a1e8... | {
"path": "tasks/E28/A/sim.mp4",
"sha256": "f82d008f455d2547227ae6fdccfd942d28a00f5ac9482d9981a9da695fb1cc03",
"bytes": 2068449,
"duration_seconds": 136.28,
"frame_count": 3407,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "390af95de7... | available | # E28A · 用平铲选取食物装盘
## 执行任务目标
按示范的食物选择方式,用平铲将相应食物送到餐盘里。
## 原始局部示范任务
按示范的食物选择方式,用平铲将相应食物送到餐盘里。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E28A 文字规则
铲起黄色长方食物放入盘中
|
E28 | B | 用平铲选取食物装盘 | tasks/E28/B/task.json | tasks/E28/B/prompt.md | tasks/E28/B/text_rule.md | {
"path": "tasks/E28/B/real.mp4",
"sha256": "a4e38034ddd5d69bc3e781a38e3e9eb9f1a85b6ee292c82121bb3ab9f29a248d",
"bytes": 3470028,
"duration_seconds": 5.9,
"frame_count": 177,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "9cf38d9a769... | {
"path": "tasks/E28/B/sim.mp4",
"sha256": "f2ec5d4249fed42fba57c3101330d5a39c042ce8bf277956ca1636e845461a23",
"bytes": 2089825,
"duration_seconds": 142.92,
"frame_count": 3573,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "6ce01004a6... | available | # E28B · 用平铲选取食物装盘
## 执行任务目标
按示范的食物选择方式,用平铲将相应食物送到餐盘里。
## 原始局部示范任务
按示范的食物选择方式,用平铲将相应食物送到餐盘里。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E28B 文字规则
铲起白黄色圆形食物放入盘中
|
E28 | C | 用平铲选取食物装盘 | tasks/E28/C/task.json | tasks/E28/C/prompt.md | tasks/E28/C/text_rule.md | {
"path": "tasks/E28/C/real.mp4",
"sha256": "09c2444e9a205a5fbc0c384f14a38c790db6ee095546990cafbec8701749e610",
"bytes": 5354240,
"duration_seconds": 8.433333,
"frame_count": 253,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "c6d3f6... | {
"path": "tasks/E28/C/sim.mp4",
"sha256": "51dd70cb725d368311553cdb363179a950baeb7d6585438daf57f70006c775b8",
"bytes": 4444823,
"duration_seconds": 295.24,
"frame_count": 7381,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "a4b5c923c1... | available | # E28C · 用平铲选取食物装盘
## 执行任务目标
按示范的食物选择方式,用平铲将相应食物送到餐盘里。
## 原始局部示范任务
按示范的食物选择方式,用平铲将相应食物送到餐盘里。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E28C 文字规则
两块食物都铲入盘中
|
E29 | A | 琴槌敲击音阶顺序 | tasks/E29/A/task.json | tasks/E29/A/prompt.md | tasks/E29/A/text_rule.md | {
"path": "tasks/E29/A/real.mp4",
"sha256": "424d3f80848e774ebf71a2172beebad88927d2e0efff3593b9b2fa9d3b993b29",
"bytes": 6056650,
"duration_seconds": 8,
"frame_count": 240,
"fps": "30/1",
"width": 1920,
"height": 1080,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "15474b98dfbfc... | {
"path": "tasks/E29/A/sim.mp4",
"sha256": "9c5f550b1fb319a2bdecfbe849bc83aa227322f85380ac42bb4724511724c437",
"bytes": 678336,
"duration_seconds": 31.56,
"frame_count": 789,
"fps": "25/1",
"width": 640,
"height": 480,
"audio_tracks": 0,
"video_codec": "h264",
"video_stream_sha256": "6f8406dd149e4... | available | # E29A · 琴槌敲击音阶顺序
## 执行任务目标
用琴槌按照示范的键位先后顺序敲击这套八键琴。
## 原始局部示范任务
用琴槌按照示范的键位先后顺序敲击这套八键琴。
逐回合实际输入以 benchmark/assignments.jsonl 中的 instruction、text_rule 和视频 SHA-256 为准。
| # E29A 文字规则
从最长键到最短键依次各敲1次:1→2→3→4→5→6→7→8
|
End of preview. Expand in Data Studio
VideoICL 具身题目与模型输入
50 题,每题 A/B/C。提供 150 段真实示范与 150 段仿真示范,全部为 H264、无音轨。每个任务目录都有文字 prompt、文字规则和任务构造信息。
| 查看内容 | 入口 |
|---|---|
| 按题目浏览 prompt、规则和 real/sim 视频 | 50 题索引 |
| 每十题一个可离线解压的包 | 五个任务包与哈希 · archives/ |
| 实际评测使用的 255 个固定物理输入与 51 个独立 QA | evaluation/ |
| 一次下载固定评测的全部视频与文字输入 | 固定输入包 |
| 各模型各回合保存的初始 prompt 和后续公开文字上下文 | model_inputs/ |
| Harness 的模型说明、系统文本与工具说明 | prompt_templates/ |
| 按模型和回合查询保存的输入 | MODEL_INPUT_INDEX.csv |
| R/I/E 环境支持信息 | environments/ |
| 学长交接、模型设置和运行入口 | GitHub 交接 |
目录
tasks/E01/A/含prompt.md、text_rule.md、task.json、real.mp4、sim.mp4,其他题目和变体同样组织。task.json记录局部 A/B/C 规则、完整任务目标、环境支持、评分设计和视频来源。任务包反映作者构造,实际运行中的逐环境指令以evaluation/cases/<episode_id>/instruction.md为准。evaluation/assignments.jsonl标明 real、sim、text_rule 或 none。视频使用消音发布版,evaluated_input_sha256追溯原评测输入。model_inputs/<model>/<experiment>/<task>/<case>/保存该物理回合的任务 prompt、公开上下文及输入覆盖说明。运行产生的 GUI 图像通过原始帧哈希关联轨迹库。evaluation/question_bank50.json是 50 道选择题,作者离线判分使用author_qa_answers.jsonl。FILE_MANIFEST.jsonl核对任务材料文件。evaluation/video_manifest.jsonl核对固定输入的视频转换。
Real 原源文件采用 HEVC,发布版转为 6Mbps H264,并核对帧数、显示方向与时长。Sim 复制原 H264 视频流并去音轨。E37 的三段仿真演示来自后补固定版本,来源见逐题元数据。
执行轨迹、原生分数、逐步动作和完整观察归档单独保存。查看这里的输入和轨迹均可匿名下载。
题目文字与 Harness 指令包 · 全部已保存模型输入文字包。
HF 查看器的 task_catalog 配置直接显示 prompt_text 和 text_context 正文;model_inputs 配置直接显示逐模型回合的 initial_prompt_text。完整上下文见记录中的文件路径。若HF查看器尚在生成,可直接打开对应md/txt文件,或下载CSV和文字包。
- Downloads last month
- 240