Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code: DatasetGenerationError
Exception: CastError
Message: Couldn't cast
id: string
source: string
split: string
lines: list<item: string>
child 0, item: string
plan: list<item: string>
child 0, item: string
replies: list<item: string>
child 0, item: string
parts: list<item: string>
child 0, item: string
complete: bool
overall: struct<low: list<item: double>, high: list<item: double>, size: list<item: double>>
child 0, low: list<item: double>
child 0, item: double
child 1, high: list<item: double>
child 0, item: double
child 2, size: list<item: double>
child 0, item: double
negative: bool
lines_by_kernel: int64
cells: list<item: string>
child 0, item: string
geom_fingerprint: string
license: string
generator_version: string
caption_style: string
caption_model: string
provenance: struct<language: string, data_version: string, plans_data: string, reader: string, resolver: string, (... 77 chars omitted)
child 0, language: string
child 1, data_version: string
child 2, plans_data: string
child 3, reader: string
child 4, resolver: string
child 5, generator: string
child 6, tier: string
child 7, stream: string
child 8, seed: int64
child 9, names: string
held_out: list<item: null>
child 0, item: null
kernel: struct<built: bool, problems: list<item: null>>
child 0, built: bool
child 1, problems: list<item: null>
child 0, item: null
to
{'id': Value('string'), 'source': Value('string'), 'split': Value('string'), 'lines': List(Value('string')), 'plan': List(Json(decode=True)), 'replies': List(Json(decode=True)), 'parts': List(Json(decode=True)), 'complete': Value('bool'), 'overall': {'low': List(Value('float64')), 'high': List(Value('float64')), 'size': List(Value('float64'))}, 'negative': Value('bool'), 'lines_by_kernel': Value('int64'), 'cells': List(Value('string')), 'geom_fingerprint': Value('string'), 'license': Value('string'), 'generator_version': Value('string'), 'caption_style': Value('string'), 'caption_model': Value('string'), 'provenance': {'language': Value('string'), 'data_version': Value('string'), 'plans_data': Value('string'), 'reader': Value('string'), 'resolver': Value('string'), 'generator': Value('string'), 'tier': Value('string'), 'stream': Value('string'), 'seed': Value('int64'), 'names': Value('string')}, 'held_out': List(Value('null'))}
because column names don't match
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
for key, table in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
id: string
source: string
split: string
lines: list<item: string>
child 0, item: string
plan: list<item: string>
child 0, item: string
replies: list<item: string>
child 0, item: string
parts: list<item: string>
child 0, item: string
complete: bool
overall: struct<low: list<item: double>, high: list<item: double>, size: list<item: double>>
child 0, low: list<item: double>
child 0, item: double
child 1, high: list<item: double>
child 0, item: double
child 2, size: list<item: double>
child 0, item: double
negative: bool
lines_by_kernel: int64
cells: list<item: string>
child 0, item: string
geom_fingerprint: string
license: string
generator_version: string
caption_style: string
caption_model: string
provenance: struct<language: string, data_version: string, plans_data: string, reader: string, resolver: string, (... 77 chars omitted)
child 0, language: string
child 1, data_version: string
child 2, plans_data: string
child 3, reader: string
child 4, resolver: string
child 5, generator: string
child 6, tier: string
child 7, stream: string
child 8, seed: int64
child 9, names: string
held_out: list<item: null>
child 0, item: null
kernel: struct<built: bool, problems: list<item: null>>
child 0, built: bool
child 1, problems: list<item: null>
child 0, item: null
to
{'id': Value('string'), 'source': Value('string'), 'split': Value('string'), 'lines': List(Value('string')), 'plan': List(Json(decode=True)), 'replies': List(Json(decode=True)), 'parts': List(Json(decode=True)), 'complete': Value('bool'), 'overall': {'low': List(Value('float64')), 'high': List(Value('float64')), 'size': List(Value('float64'))}, 'negative': Value('bool'), 'lines_by_kernel': Value('int64'), 'cells': List(Value('string')), 'geom_fingerprint': Value('string'), 'license': Value('string'), 'generator_version': Value('string'), 'caption_style': Value('string'), 'caption_model': Value('string'), 'provenance': {'language': Value('string'), 'data_version': Value('string'), 'plans_data': Value('string'), 'reader': Value('string'), 'resolver': Value('string'), 'generator': Value('string'), 'tier': Value('string'), 'stream': Value('string'), 'seed': Value('int64'), 'names': Value('string')}, 'held_out': List(Value('null'))}
because column names don't match
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/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1879, in _prepare_split_single
raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the datasetNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
id string | source string | split string | lines list | plan list | replies list | parts list | complete bool | overall dict | negative bool | lines_by_kernel int64 | cells list | geom_fingerprint string | license string | generator_version string | caption_style string | caption_model string | provenance dict | held_out list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d29c2cc5e46e816d | negatives | train | [
"shelf: box 440 by 120 by 60, on ground",
"group skid",
" cleat: cylinder 40 by 40 lying along depth",
" step: box 5 by 50 by 20 lying along depth, under cleat",
" gusset 5: box 15 by 25 by 20, in front of cleat",
" fd22: cylinder 12 by same as step's length lying along depth, left of gusset 5, top at... | [
{
"number": 1,
"text": "shelf: box 440 by 120 by 60, on ground",
"kind": "part",
"name": "shelf",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 440
},
{
"slot": "depth",
"source": "number",
"value... | [
{
"line": 1,
"reply": "built",
"echo": "shelf: box 440 by 120 by 60, on ground",
"made": [
"shelf"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "group skid",
"by_kernel": false
},
{
"line": 3,
"reply": "built",
"echo": "cleat: cyli... | [
{
"name": "shelf",
"owner": "shelf",
"line": 1,
"shape": "box",
"sizes": {
"length": 440,
"depth": 120,
"height": 60
},
"low": [
-220,
-60,
0
],
"high": [
220,
60,
60
],
"size": [
440,
120,
60
],
... | false | {
"low": [
-220,
-60,
0
],
"high": [
260,
60,
90
],
"size": [
480,
120,
90
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"group:declared",
"reply:built",
"shape:cylinder|lying along depth",
"size:number",
"size:number",
"group:first part",
"reply:built",
"s... | n6|b90.0x120.0x480.0|v3231896.46|5b7a1957ff | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 0,
"names": "template"
} | [] |
ea75ed4f4c218d31 | negatives | train | [
"ua2: box 180 by 120 by 60, on ground",
"kv40: box 60 by 25 by 30 lying along depth, in front of ua2",
"gusset: box 25.5 by 15 by 12 lying along depth, under kv40",
"wc27: cylinder 12 by rest lying along depth, between ua2 and gusset",
"band: cylinder 30 by 70 lying along length, on ground, left flush with ... | [
{
"number": 1,
"text": "ua2: box 180 by 120 by 60, on ground",
"kind": "part",
"name": "ua2",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 180
},
{
"slot": "depth",
"source": "number",
"value": 1... | [
{
"line": 1,
"reply": "built",
"echo": "ua2: box 180 by 120 by 60, on ground",
"made": [
"ua2"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "kv40: box 60 by 25 by 30 lying along depth, in front of ua2",
"made": [
"kv40"
],
"by_kernel... | [
{
"name": "ua2",
"owner": "ua2",
"line": 1,
"shape": "box",
"sizes": {
"length": 180,
"depth": 120,
"height": 60
},
"low": [
-90,
-60,
0
],
"high": [
90,
60,
60
],
"size": [
180,
120,
60
],
"c... | false | {
"low": [
-90,
-90,
0
],
"high": [
160,
81,
150
],
"size": [
250,
171,
150
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"shape:box|lying along depth",
"size:number",
"size:number",
"size:number",
"place:in front of",
"pair:shape×place|box|in front of",
"pair... | n10|b150.0x171.0x250.0|v1450952.95|3dd740b1d5 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 1,
"names": "template"
} | [] |
9bad7dfea8d509b5 | negatives | train | [
"plug: box 180 by 100 by 60 lying along depth, on ground",
"tile: box 40 by 40 by 40, left of plug",
"stud: box same as tile's length by 20 by 60 lying along length, right of plug",
"rn33: box same as tile's length by 20 by 60 lying along length, right of plug",
"strut: box 40 by 100 by default lying along ... | [
{
"number": 1,
"text": "plug: box 180 by 100 by 60 lying along depth, on ground",
"kind": "part",
"name": "plug",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 180
},
{
"slot": "depth",
"source": "number"... | [
{
"line": 1,
"reply": "built",
"echo": "plug: box 180 by 100 by 60 lying along depth, on ground",
"made": [
"plug"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "tile: box 40 by 40 by 40, left of plug",
"made": [
"tile"
],
"by_kernel"... | [
{
"name": "plug",
"owner": "plug",
"line": 1,
"shape": "box",
"sizes": {
"length": 180,
"depth": 100,
"height": 60
},
"low": [
-90,
-30,
0
],
"high": [
90,
30,
100
],
"size": [
180,
60,
100
],
... | false | {
"low": [
-130,
-30,
0
],
"high": [
150,
70,
200
],
"size": [
280,
100,
200
]
} | true | 0 | [
"reply:built",
"shape:box|lying along depth",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:left of",
"pair:shape×place|box|left of",
"pair:place×t... | n12|b100.0x200.0x280.0|v1403718.28|04e641dc67 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 2,
"names": "template"
} | [] |
7c324becfe0d461e | negatives | train | [
"pillar: box 520 by 440 by 10, on ground",
"long arms: box 20 by 5 by 50 lying along depth, behind pillar, 2 around pillar facing outward",
"step: box 25 by 60 by half of long arms' depth lying along depth, on top of pillar, front flush with pillar's back, flush with pillar's back, inset 25",
"bf20: box 50 by... | [
{
"number": 1,
"text": "pillar: box 520 by 440 by 10, on ground",
"kind": "part",
"name": "pillar",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 520
},
{
"slot": "depth",
"source": "number",
"val... | [
{
"line": 1,
"reply": "built",
"echo": "pillar: box 520 by 440 by 10, on ground",
"made": [
"pillar"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "long arms: box 20 by 5 by 50 lying along depth, behind pillar, 2 around pillar facing outward",
"mad... | [
{
"name": "pillar",
"owner": "pillar",
"line": 1,
"shape": "box",
"sizes": {
"length": 520,
"depth": 440,
"height": 10
},
"low": [
-260,
-220,
0
],
"high": [
260,
220,
10
],
"size": [
520,
440,
10
... | false | {
"low": [
-260,
-270,
0
],
"high": [
260,
270,
205
],
"size": [
520,
540,
205
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"shape:box|lying along depth",
"size:number",
"size:number",
"size:number",
"place:behind",
"pair:shape×place|box|behind",
"pair:place×tar... | n5|b205.0x520.0x540.0|v2874766.58|8038e55597 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 3,
"names": "template"
} | [] |
bc15ae6cb34aee24 | negatives | train | [
"web: box 460 by 240 by 40, on ground",
"group wing 2",
" sleeve: box 60 by 40 by 40 lying along length",
" lid: box 8 by 80 by 30 lying along length, in front of sleeve",
" cross cap: box 120 by 8 by 5, left of lid",
"end",
"eu8: wing 2, on top of web, 2 around web on circle 180",
"wu4: cylinder 4... | [
{
"number": 1,
"text": "web: box 460 by 240 by 40, on ground",
"kind": "part",
"name": "web",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 460
},
{
"slot": "depth",
"source": "number",
"value": 2... | [
{
"line": 1,
"reply": "built",
"echo": "web: box 460 by 240 by 40, on ground",
"made": [
"web"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "group wing 2",
"by_kernel": false
},
{
"line": 3,
"reply": "built",
"echo": "sleeve: box 6... | [
{
"name": "web",
"owner": "web",
"line": 1,
"shape": "box",
"sizes": {
"length": 460,
"depth": 240,
"height": 40
},
"low": [
-230,
-120,
0
],
"high": [
230,
120,
40
],
"size": [
460,
240,
40
],
... | false | {
"low": [
-230,
-260,
0
],
"high": [
230,
205,
360
],
"size": [
460,
465,
360
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"group:declared",
"reply:built",
"shape:box|lying along length",
"size:number",
"size:number",
"size:number",
"group:first part",
"reply... | n28|b360.0x460.0x465.0|v6901888.22|13a645b951 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 4,
"names": "template"
} | [] |
fafb93f91fcd0464 | negatives | train | [
"end cap: box 540 by 260 by 10, on ground",
"ledge: cylinder 5 by 40 lying along length, behind end cap",
"leg: box 40.5 by 150 by 60, above ground, bottom at height 300, left 40 beyond end cap's left",
"th40: box rest by 12 by 12 lying along length, under leg, down to ground",
"tie: box 30 by 8 by rest, on... | [
{
"number": 1,
"text": "end cap: box 540 by 260 by 10, on ground",
"kind": "part",
"name": "end cap",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 540
},
{
"slot": "depth",
"source": "number",
"v... | [
{
"line": 1,
"reply": "built",
"echo": "end cap: box 540 by 260 by 10, on ground",
"made": [
"end cap"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "ledge: cylinder 5 by 40 lying along length, behind end cap",
"made": [
"ledge"
],
"b... | [
{
"name": "end cap",
"owner": "end cap",
"line": 1,
"shape": "box",
"sizes": {
"length": 540,
"depth": 260,
"height": 10
},
"low": [
-270,
-130,
0
],
"high": [
270,
130,
10
],
"size": [
540,
260,
10
... | false | {
"low": [
-349.166667,
-135,
0
],
"high": [
270,
265,
475
],
"size": [
619.166667,
400,
475
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"shape:cylinder|lying along length",
"size:number",
"size:number",
"place:behind",
"pair:shape×place|cylinder|behind",
"pair:place×target|be... | n24|b400.0x475.0x619.2|v3875571.54|c2cb35f4a3 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 5,
"names": "template"
} | [] |
5b0d4e57ddd991c1 | negatives | train | [
"base: box 420 by 320 by 10, on ground",
"lower tie: cylinder 20 by 70 lying along depth, on ground, front flush with base's back",
"pin: box 30 by 70 by 20, right of lower tie",
"decks: box 5 by 100 by 5 lying along length, behind lower tie, flush with lower tie's bottom, inset 30, grid 2 by 1",
"end guide... | [
{
"number": 1,
"text": "base: box 420 by 320 by 10, on ground",
"kind": "part",
"name": "base",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 420
},
{
"slot": "depth",
"source": "number",
"value":... | [
{
"line": 1,
"reply": "built",
"echo": "base: box 420 by 320 by 10, on ground",
"made": [
"base"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "lower tie: cylinder 20 by 70 lying along depth, on ground, front flush with base's back",
"made": [
... | [
{
"name": "base",
"owner": "base",
"line": 1,
"shape": "box",
"sizes": {
"length": 420,
"depth": 320,
"height": 10
},
"low": [
-210,
-160,
0
],
"high": [
210,
160,
10
],
"size": [
420,
320,
10
],
... | false | {
"low": [
-210,
-300,
0
],
"high": [
265,
330,
62.5
],
"size": [
475,
630,
62.5
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"shape:cylinder|lying along depth",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|cylinder|on ground",
"align:flush-opp... | n15|b62.5x475.0x630.0|v1502374.94|382791ada8 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 6,
"names": "template"
} | [] |
595fb56eed0db582 | negatives | train | [
"end leg: box 520 by 420 by 10 standing, above ground, top at height 110",
"keys 8: box 50 by 50 by rest, under end leg, down to ground, at the left corners of end leg",
"group brace",
" knob: cylinder 20 by 25",
" tab: cylinder 20 by 80, under knob",
"end",
"small collars: brace, in front of end leg,... | [
{
"number": 1,
"text": "end leg: box 520 by 420 by 10 standing, above ground, top at height 110",
"kind": "part",
"name": "end leg",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 520
},
{
"slot": "depth",
... | [
{
"line": 1,
"reply": "built",
"echo": "end leg: box 520 by 420 by 10, above ground, top at height 110",
"made": [
"end leg"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "keys 8: box 50 by 50 by rest, under end leg, down to ground, at the left corners... | [
{
"name": "end leg",
"owner": "end leg",
"line": 1,
"shape": "box",
"sizes": {
"length": 520,
"depth": 420,
"height": 10
},
"low": [
-260,
-210,
100
],
"high": [
260,
210,
110
],
"size": [
520,
420,
1... | false | {
"low": [
-260,
-310,
0
],
"high": [
332.5,
210,
265
],
"size": [
592.5,
520,
265
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:above ground|top at height",
"pair:shape×place|box|above ground",
"align:top at height",
"pair:place×align|above ground|at height",
"amount:number",
"reply:built",
"shape:box|standing",
"size:number",
... | n46|b265.0x520.0x592.5|v3769986.69|fb687376c2 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 9,
"names": "template"
} | [] |
5b6c389345b8e9ab | negatives | train | [
"small base: box 210 by 70 by 90 lying along length, on ground",
"pad 3: cylinder 5 by 30, in front of small base",
"panel 7: cylinder 5 by 50 lying along length, right of pad 3, gap 25, front flush with small base's back, front 40 beyond pad 3's front",
"deck: box rest by 10 by 10, between panel 7 and pad 3"... | [
{
"number": 1,
"text": "small base: box 210 by 70 by 90 lying along length, on ground",
"kind": "part",
"name": "small base",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 210
},
{
"slot": "depth",
"sourc... | [
{
"line": 1,
"reply": "built",
"echo": "small base: box 210 by 70 by 90 lying along length, on ground",
"made": [
"small base"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "pad 3: cylinder 5 by 30, in front of small base",
"made": [
"pad 3"
... | [
{
"name": "small base",
"owner": "small base",
"line": 1,
"shape": "box",
"sizes": {
"length": 210,
"depth": 70,
"height": 90
},
"low": [
-45,
-35,
0
],
"high": [
45,
35,
210
],
"size": [
90,
70,
210
... | false | {
"low": [
-92.5,
-80,
0
],
"high": [
82.5,
35,
210
],
"size": [
175,
115,
210
]
} | true | 0 | [
"reply:built",
"shape:box|lying along length",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"shape:cylinder|standing",
"size:number",
"size:number",
"place:in front of",
"pair:shape×place|cylinder|in front of",
"pair:place... | n10|b115.0x175.0x210.0|v1466026.85|ee229178ad | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 10,
"names": "template"
} | [] |
ac5af28dd4752fbb | negatives | train | [
"ee9: box 260 by 480 by 30, on ground",
"group column",
" key: box 40 by 80 by 25 lying along depth",
" rail: box 40 by 20 by 40, right of key",
"end",
"pin: column, on top of ee9",
"cleats: column, behind pin, down to ground, flush with pin's right, inset 15, mirrored left-right",
"lower strip: col... | [
{
"number": 1,
"text": "ee9: box 260 by 480 by 30, on ground",
"kind": "part",
"name": "ee9",
"shape": "box",
"sizes": [
{
"slot": "length",
"source": "number",
"value": 260
},
{
"slot": "depth",
"source": "number",
"value": 4... | [
{
"line": 1,
"reply": "built",
"echo": "ee9: box 260 by 480 by 30, on ground",
"made": [
"ee9"
],
"by_kernel": false
},
{
"line": 2,
"reply": "built",
"echo": "group column",
"by_kernel": false
},
{
"line": 3,
"reply": "built",
"echo": "key: box 40 b... | [
{
"name": "ee9",
"owner": "ee9",
"line": 1,
"shape": "box",
"sizes": {
"length": 260,
"depth": 480,
"height": 30
},
"low": [
-130,
-240,
0
],
"high": [
130,
240,
30
],
"size": [
260,
480,
30
],
... | false | {
"low": [
-130,
-240,
0
],
"high": [
130,
320,
110
],
"size": [
260,
560,
110
]
} | true | 0 | [
"reply:built",
"shape:box|standing",
"size:number",
"size:number",
"size:number",
"place:on ground",
"pair:shape×place|box|on ground",
"reply:built",
"group:declared",
"reply:built",
"shape:box|lying along depth",
"size:number",
"size:number",
"size:number",
"group:first part",
"reply:... | n11|b110.0x260.0x560.0|v4257185.01|cff50682e2 | forge | cf0f2ca923af | plan | template | {
"language": "docs/PLAN_LANGUAGE.md Draft 2",
"data_version": "plans-v1",
"plans_data": "cf0f2ca923af",
"reader": "04a9265155a8",
"resolver": "53c56a3fba75",
"generator": "random_gen-1",
"tier": "arithmetic",
"stream": "negatives-arithmetic",
"seed": 11,
"names": "template"
} | [] |
End of preview.
Forge-D1
Training and test data for Forge-S1, a small step-by-step CAD executor. At each step the model reads a structured plan and the live state of a FreeCAD session, and picks the next command.
Every session here was recorded in real FreeCAD (headless) by a scripted teacher, with wrong commands injected on purpose so that the data shows how to notice and undo a mistake. The label at each step is the set of commands the teacher accepts; an injected mistake is never a label.
All of it was produced by the project's own generator code. source and generator_version are on
every row.
Code: https://github.com/yrevash/forge
Contents
| Folder | Contents | State |
|---|---|---|
single_part/ |
363,471 sessions, 19,666,441 labelled steps | Verified: 2,409 sessions replayed in FreeCAD with no label difference |
structures_v2_unaudited/ |
61,057 multi-part sessions, 10,593,975 steps | Not yet verified. Do not train on it. Test slices complete; train is 2,027 of 4,781 files |
plans/ |
271,623 plans in the plan language, with resolved parts | About 1 in 150 random plans with copied groups has overlapping parts |
structures/ |
52,684 verified structures of 18 kinds, with prompts | Verified, 0 failures |
testplans_v1_TEST_ONLY/ |
120 test requests with a plan for each | Test only. Never train on it |
single_part/ slices
| Slice | Mistake mix | Sessions | Steps | Use |
|---|---|---|---|---|
train_v2 |
improved | 97,971 | 5,575,981 | main training slice |
train |
first | 97,971 | 4,559,409 | training |
train_s1 |
first | 97,971 | 4,570,035 | second session of every train part |
train_s2 |
first | 14,500 | 676,519 | third session, partial |
train_heavy |
first, noise 0.4 / 0.5 | 3,000 | 348,322 | heavy noise; keep out of a first training mix |
iid_v2, pairing_v2, long_v2 |
improved | 5,133 / 16,896 / 4,000 | 289,811 / 1,285,181 / 620,989 | tests: like training, unseen pairings, long plans |
iid, pairing, long |
first | 5,133 / 16,896 / 4,000 | 237,896 / 1,038,184 / 464,114 | the same test parts with the first mix |
- The improved mix adds:
donesaid too early, features that fail to build, a lost undo history, and more wrong-number mistakes (near misses such as swapped length and width). - Two sessions are excluded from use:
8faf9b4de268c2bdand118d531507fc2bc2(both ended abnormally). - 63% of
longparts also contain a held-out pairing;long_pure.jsonlists the ones that do not. - No plan and no shape is shared between a train slice and a test slice.
File format
Each slice is a folder of shardNNN.jsonl.gz files with a shardNNN.stats.json beside each.
A shard holds whole sessions. Per session: one header line (session id, part id, split, slice, seed,
noise level, start state, the plan), then one line per step (the snapshot of the FreeCAD session,
the valid commands, the teacher's target set, what was executed and FreeCAD's reply), then an end line.
Do not feed raw records to a model. The header and the step records hold label-side fields (noise
level, which command was the injected mistake, progress). Read them only through
forge/freecad/load.py in the Forge code, which passes the plan, the snapshot and the valid
commands to the model and the target set as the label:
from pathlib import Path
from forge.freecad import load
for example in load.examples(["train_v2"], out_dir=Path("single_part")):
example.view # what the model may see
example.label # the set of right answers
The floor a model must beat
A rule that never reads the plan (fitted on train_v2, single_part/baselines_v2.log) gets about
68% of all steps right on the improved tests, but only 25–27% of undo decisions, 33–35% of steps
right after a mistake, and 4–6% of wrong-number mistakes. Report step accuracy beside this floor.
The headline measure is whether the finished part is correct when the model drives FreeCAD itself.
Download one slice
hf download yrevash/forge-d1 --repo-type=dataset --include "single_part/train_v2/*" --local-dir forge-d1
Licence and credit
Creative Commons Attribution 4.0 (CC BY 4.0). You may use, change and share it, also commercially. The one condition is credit: say "Forge-S1 by Yash Tiwari (yrevash)" and link to https://github.com/yrevash/forge.
- Downloads last month
- 81