LAGEN
Collection
Models, datasets, and latency profiles for LAGEN. • 3 items • Updated
Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
burst_dwell_distribution: struct<distribution_type: string, latency_ms: list<item: double>, quantile_levels: list<item: double (... 2 chars omitted)
child 0, distribution_type: string
child 1, latency_ms: list<item: double>
child 0, item: double
child 2, quantile_levels: list<item: double>
child 0, item: double
burst_dwell_lengths: list<item: int64>
child 0, item: int64
burst_merge_gap_records: int64
burst_rank_processes: list<item: struct<draw_count: int64, dwell_length_spearman_rho: double, level_ms: struct<distributio (... 343 chars omitted)
child 0, item: struct<draw_count: int64, dwell_length_spearman_rho: double, level_ms: struct<distribution_type: str (... 331 chars omitted)
child 0, draw_count: int64
child 1, dwell_length_spearman_rho: double
child 2, level_ms: struct<distribution_type: string, latency_ms: list<item: double>, quantile_levels: list<item: double (... 2 chars omitted)
child 0, distribution_type: string
child 1, latency_ms: list<item: double>
child 0, item: double
child 2, quantile_levels: list<item: double>
child 0, item: double
child 3, level_residual_ms: struct<distribution_type: string, latency_ms: list<item: double>, quantile_levels: list<item: double (... 2 chars omitted)
child 0, distribution_type: string
child 1, latency_ms: list<item: double>
child 0, item: double
child 2, quantile_levels: list<item:
...
ol, cuda_math_sdp_en (... 447 chars omitted)
child 0, backends: struct<cuda_cudnn_sdp_enabled: bool, cuda_flash_sdp_enabled: bool, cuda_math_sdp_enabled: bool, cuda (... 110 chars omitted)
child 0, cuda_cudnn_sdp_enabled: bool
child 1, cuda_flash_sdp_enabled: bool
child 2, cuda_math_sdp_enabled: bool
child 3, cuda_matmul_allow_tf32: bool
child 4, cuda_mem_efficient_sdp_enabled: bool
child 5, cudnn_allow_tf32: bool
child 6, cudnn_benchmark: bool
child 1, cuda_available: bool
child 2, cuda_device_count: int64
child 3, cuda_version: string
child 4, current_device: int64
child 5, current_device_name: string
child 6, device_properties: list<item: struct<index: int64, major: int64, minor: int64, multi_processor_count: int64, name: stri (... 25 chars omitted)
child 0, item: struct<index: int64, major: int64, minor: int64, multi_processor_count: int64, name: string, total_m (... 13 chars omitted)
child 0, index: int64
child 1, major: int64
child 2, minor: int64
child 3, multi_processor_count: int64
child 4, name: string
child 5, total_memory: int64
child 7, float32_matmul_precision: string
child 8, version: string
instance_id: string
gpus: list<item: struct<name: string, slot: int64>>
child 0, item: struct<name: string, slot: int64>
child 0, name: string
child 1, slot: int64
topology_links: list<item: null>
child 0, item: null
gpu_class: string
to
{'driver_version': Value('string'), 'gpu_class': Value('string'), 'gpus': List({'name': Value('string'), 'slot': Value('int64')}), 'instance_id': Value('string'), 'topology_links': List(Value('null')), 'torch': {'backends': {'cuda_cudnn_sdp_enabled': Value('bool'), 'cuda_flash_sdp_enabled': Value('bool'), 'cuda_math_sdp_enabled': Value('bool'), 'cuda_matmul_allow_tf32': Value('bool'), 'cuda_mem_efficient_sdp_enabled': Value('bool'), 'cudnn_allow_tf32': Value('bool'), 'cudnn_benchmark': Value('bool')}, 'cuda_available': Value('bool'), 'cuda_device_count': Value('int64'), 'cuda_version': Value('string'), 'current_device': Value('int64'), 'current_device_name': Value('string'), 'device_properties': List({'index': Value('int64'), 'major': Value('int64'), 'minor': Value('int64'), 'multi_processor_count': Value('int64'), 'name': Value('string'), 'total_memory': Value('int64')}), 'float32_matmul_precision': Value('string'), 'version': Value('string')}}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
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
burst_dwell_distribution: struct<distribution_type: string, latency_ms: list<item: double>, quantile_levels: list<item: double (... 2 chars omitted)
child 0, distribution_type: string
child 1, latency_ms: list<item: double>
child 0, item: double
child 2, quantile_levels: list<item: double>
child 0, item: double
burst_dwell_lengths: list<item: int64>
child 0, item: int64
burst_merge_gap_records: int64
burst_rank_processes: list<item: struct<draw_count: int64, dwell_length_spearman_rho: double, level_ms: struct<distributio (... 343 chars omitted)
child 0, item: struct<draw_count: int64, dwell_length_spearman_rho: double, level_ms: struct<distribution_type: str (... 331 chars omitted)
child 0, draw_count: int64
child 1, dwell_length_spearman_rho: double
child 2, level_ms: struct<distribution_type: string, latency_ms: list<item: double>, quantile_levels: list<item: double (... 2 chars omitted)
child 0, distribution_type: string
child 1, latency_ms: list<item: double>
child 0, item: double
child 2, quantile_levels: list<item: double>
child 0, item: double
child 3, level_residual_ms: struct<distribution_type: string, latency_ms: list<item: double>, quantile_levels: list<item: double (... 2 chars omitted)
child 0, distribution_type: string
child 1, latency_ms: list<item: double>
child 0, item: double
child 2, quantile_levels: list<item:
...
ol, cuda_math_sdp_en (... 447 chars omitted)
child 0, backends: struct<cuda_cudnn_sdp_enabled: bool, cuda_flash_sdp_enabled: bool, cuda_math_sdp_enabled: bool, cuda (... 110 chars omitted)
child 0, cuda_cudnn_sdp_enabled: bool
child 1, cuda_flash_sdp_enabled: bool
child 2, cuda_math_sdp_enabled: bool
child 3, cuda_matmul_allow_tf32: bool
child 4, cuda_mem_efficient_sdp_enabled: bool
child 5, cudnn_allow_tf32: bool
child 6, cudnn_benchmark: bool
child 1, cuda_available: bool
child 2, cuda_device_count: int64
child 3, cuda_version: string
child 4, current_device: int64
child 5, current_device_name: string
child 6, device_properties: list<item: struct<index: int64, major: int64, minor: int64, multi_processor_count: int64, name: stri (... 25 chars omitted)
child 0, item: struct<index: int64, major: int64, minor: int64, multi_processor_count: int64, name: string, total_m (... 13 chars omitted)
child 0, index: int64
child 1, major: int64
child 2, minor: int64
child 3, multi_processor_count: int64
child 4, name: string
child 5, total_memory: int64
child 7, float32_matmul_precision: string
child 8, version: string
instance_id: string
gpus: list<item: struct<name: string, slot: int64>>
child 0, item: struct<name: string, slot: int64>
child 0, name: string
child 1, slot: int64
topology_links: list<item: null>
child 0, item: null
gpu_class: string
to
{'driver_version': Value('string'), 'gpu_class': Value('string'), 'gpus': List({'name': Value('string'), 'slot': Value('int64')}), 'instance_id': Value('string'), 'topology_links': List(Value('null')), 'torch': {'backends': {'cuda_cudnn_sdp_enabled': Value('bool'), 'cuda_flash_sdp_enabled': Value('bool'), 'cuda_math_sdp_enabled': Value('bool'), 'cuda_matmul_allow_tf32': Value('bool'), 'cuda_mem_efficient_sdp_enabled': Value('bool'), 'cudnn_allow_tf32': Value('bool'), 'cudnn_benchmark': Value('bool')}, 'cuda_available': Value('bool'), 'cuda_device_count': Value('int64'), 'cuda_version': Value('string'), 'current_device': Value('int64'), 'current_device_name': Value('string'), 'device_properties': List({'index': Value('int64'), 'major': Value('int64'), 'minor': Value('int64'), 'multi_processor_count': Value('int64'), 'name': Value('string'), 'total_memory': Value('int64')}), 'float32_matmul_precision': Value('string'), 'version': Value('string')}}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
LAGEN latency profiles
Project resources: LAGEN collection.
Measured latency profiles, fitted distributions and measurement evidence.
| Experiment | Entry |
|---|---|
| Sim2Real calibration | 30-case held-out calibration |
| Visual history | Visual history |
| Latency in prompt | Latency in prompt |
| Latency transfer | Latency transfer |
| Task transfer | Task transfer |
| VLA fine-tuning scope | VLA fine-tuning scope |
| Mean vs. profile training | Mean vs. profile training |
| Observation stride | Observation stride |
| Context window | Context window |
| Resource | Repository |
|---|---|
| benchmark-models | MLL-Lab/LAGEN-models |
| benchmark-datasets | MLL-Lab/LAGEN-datasets |
Artifact retention
HAIC and Extreme Parkour releases are retired. Model bundles retain the published evaluation checkpoint, or the latest checkpoint when no evaluation selection exists. Optimizer and trainer recovery state are not release assets. Identical dataset copies use the canonical task paths. Original configurations and experiment evidence remain source records.
Historical Standard-Pipeline runs
Preserved run evidence retains native profiling directories and the original fixed revisions. Source index records exact ownership and source identities.
- Downloads last month
- 1,625