Datasets:

HanLinqi
/

robotwin-trace-points-300-20260930

Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
task: string
episode: int64
frame: int64
window: int64
split: string
hdf5: string
instruction: string
object_labels: string
ee_labels: string
unseen: list<item: string>
  child 0, item: string
seen: list<item: string>
  child 0, item: string
to
{'seen': List(Value('string')), 'unseen': List(Value('string'))}
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
              task: string
              episode: int64
              frame: int64
              window: int64
              split: string
              hdf5: string
              instruction: string
              object_labels: string
              ee_labels: string
              unseen: list<item: string>
                child 0, item: string
              seen: list<item: string>
                child 0, item: string
              to
              {'seen': List(Value('string')), 'unseen': List(Value('string'))}
              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 1880, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

seen list	unseen list
[ "Pick the bottle with ridges near base head-up using the right arm", "Lift the medium green bottle with capped neck from the table without mentioning the arm.", "Pick the bottle with rounded neck upright from the table", "Use the left arm to pick up the smooth green plastic bottle from the table.", "Use the...	[ "Pick up the green plastic bottle with ridged bottom using the left arm in an upright position", "Pick up the green plastic bottle with ridged bottom using the correct arm", "Use the left arm to grab the green bottle with white label upright", "Pick up the green plastic bottle with ridged bottom carefully usi...
[ "Raise the plastic drink bottle from the table using the correct arm, the left arm.", "Grab the bottle with ridges on bottom from the table and hold it head-up.", "Pick the dark drink bottle with logo and keep it upright", "Carefully grab the bottle with white neck and red top head-up", "Lift the medium dar...	[ "Use the left arm to grab the bottle with red screw cap upright", "Use the left arm to grab the plastic soda bottle head-up.", "Grab the bottle with red screw cap on the table and raise it.", "Grab the brown bottle from the table with the left arm", "Use the left arm to grab the bottle the bottle with red s...
[ "Hold the bottle with ridges near base upright from the table", "Grab the bottle with rounded neck and lift it with the left arm", "Pick the smooth green plastic bottle and keep it upright", "Grab the medium green bottle with capped neck and lift it upward from the table.", "Use the left arm to lift the med...	[ "Grab the green plastic bottle with ridged bottom on the table and raise it.", "Grab the green bottle with white label from the table with the left arm", "Pick up the green plastic bottle with ridged bottom carefully using the left arm", "Lift the green plastic bottle with ridged bottom from the table using t...
[ "Raise the medium dark brown bottle from the table using the correct arm, the left arm.", "Use the left arm to hold the smooth bottle with red lid from the table", "Lift the plastic drink bottle from the table with the left arm", "Lift the medium dark brown bottle off the table and hold it upright.", "Use t...	[ "Pick up the bottle with red screw cap using the correct arm", "Lift the plastic soda bottle from the table using the left arm.", "Grab the brown bottle from the table with the left arm", "Use the left arm to grab the bottle the brown bottle", "Use the left arm to grab the brown bottle head-up.", "Pick up...
[ "Pick the bottle with ridges on bottom upright from the table", "Grab the Coca-Cola bottle from the table and hold it head-up.", "Hold the narrow-necked bottle upright from the table", "Pick the brown bottle with red label head-up using the right arm", "Identify the bottle with white neck and red top and gr...	[ "Pick up the plastic soda bottle carefully using the right arm", "Use the right arm to grab the bottle with red screw cap upright", "Pick up the brown bottle using the correct arm", "Use the right arm to grab the bottle the brown bottle", "Pick up the bottle with red screw cap from the table head-up.", "P...
[ "Raise the medium green bottle with capped neck in a head-up position", "Lift the medium green bottle with capped neck using the left arm and hold upright", "Grab the bottle with rounded neck and lift it upward from the table.", "Lift the bottle with narrow top off the table and hold it upright.", "Grab the...	[ "Pick up the green bottle carefully using the left arm", "Pick up the green bottle using the left arm in an upright position", "Lift the green plastic bottle with ridged bottom from the table using the left arm.", "Grab the green plastic bottle with ridged bottom on the table and raise it.", "Pick up the gr...
[ "Grab the hand-sized soda bottle from the table with the left arm", "Use the left arm to pick up the brown bottle with red label in the correct orientation", "Pick up the hand-sized soda bottle upright from the table", "Carefully grab the Coca-Cola bottle head-up", "Use the left arm to pick up the bottle wi...	[ "Grab the bottle with red screw cap on the table and raise it.", "Pick up the plastic soda bottle using the correct arm", "Use the left arm to grab the bottle with red screw cap upright", "Pick up the brown bottle using the left arm in an upright position", "Grab the brown bottle from the table with the lef...
[ "Pick the smooth green bottle upright from the table", "Lift the smooth green plastic bottle head-up from the table", "Lift the bottle the medium green bottle with capped neck up from the table", "Lift the green bottle with indented base from the table upright", "Grab the bottle with ridges near base head-u...	[ "Grab the green bottle from the table with the right arm", "Lift the green bottle from the table using the right arm.", "Use the right arm to grab the green bottle with white label upright", "Grab the green bottle with white label on the table and raise it.", "Pick up the green bottle with white label caref...
[ "Grab the medium green bottle with capped neck and lift it with the left arm", "Locate the bottle with white printed label and lift it upright with the left arm.", "Grab the bottle with rounded neck from the table and hold it head-up.", "Carefully grab the medium green bottle head-up", "Pick up the smooth g...	[ "Pick up the green bottle with white label carefully using the left arm", "Pick up the green plastic bottle with ridged bottom using the correct arm", "Grab the green plastic bottle with ridged bottom on the table and raise it.", "Use the left arm to grab the green bottle with white label upright", "Use the...
[ "Grab the bottle with ridges on bottom from the table with the left arm", "Lift the plastic drink bottle from the table and secure upright", "Pick the Coca-Cola bottle up with the left arm ensuring it stays head-up.", "Grab the long thin bottle and lift it into a head-up position", "Use the correct arm to p...	[ "Lift the plastic soda bottle from the table using the left arm.", "Use the left arm to grab the plastic soda bottle upright", "Use the left arm to grab the bottle the plastic soda bottle", "Pick up the brown bottle using the left arm in an upright position", "Grab the plastic soda bottle on the table and r...
[ "Use the correct arm to pick up the bottle with ridges on bottom.", "Pick the narrow-necked bottle and keep it upright", "Pick the medium dark brown bottle upright from the table", "Use the right arm to pick the narrow-necked bottle up and keep it head-up.", "Lift the plastic drink bottle head-up from the t...	[ "Grab the plastic soda bottle from the table with the right arm", "Use the right arm to grab the bottle with red screw cap head-up.", "Use the right arm to grab the plastic soda bottle upright", "Lift the plastic soda bottle from the table using the right arm.", "Use the right arm to grab the bottle the bro...
[ "Pick up the medium dark brown bottle upright from the table", "Pick up the narrow-necked bottle from the table using the left arm.", "Use the left arm to lift the smooth bottle with red lid head-up", "Pick the long thin bottle up with the left arm ensuring it stays head-up.", "Find the long thin bottle on ...	[ "Lift the plastic soda bottle from the table using the left arm.", "Grab the brown bottle on the table and raise it.", "Pick up the bottle with red screw cap using the correct arm", "Use the left arm to grab the brown bottle upright", "Use the left arm to grab the plastic soda bottle head-up.", "Pick up t...
[ "Use the left arm to pick the hand-sized soda bottle up and keep it head-up.", "Pick the plastic drink bottle upright from the table", "Grab the narrow-necked bottle and lift it with the left arm", "Lift the brown bottle with red label off the table and hold it upright.", "Hold the hand-sized soda bottle he...	[ "Use the left arm to grab the bottle the brown bottle", "Grab the plastic soda bottle from the table with the left arm", "Pick up the bottle with red screw cap carefully using the left arm", "Pick up the brown bottle using the left arm in an upright position", "Pick up the bottle with red screw cap using th...
[ "Find the narrow-necked bottle on the table and raise it using the right arm.", "Lift the long thin bottle from the table without mentioning the arm.", "Hold the bottle with white neck and red top upright after lifting with the right arm", "Pick the medium dark brown bottle and keep it upright", "Lift the b...	[ "Use the right arm to grab the bottle the bottle with red screw cap", "Grab the bottle with red screw cap on the table and raise it.", "Lift the plastic soda bottle from the table using the right arm.", "Use the right arm to grab the plastic soda bottle upright", "Pick up the brown bottle from the table hea...

End of preview.

RoboTwin 300: object and end-effector point trajectories

This dataset contains 300 simulated expert episodes across six RoboTwin tasks, with 15,455 windows for predicting the motion of object and robot end-effector points. It was prepared for the Trace-MoT experiments. The data is simulation-only. The object queries use simulator segmentation to identify the intended object; this provides oracle target information.

raw/<task>/episode<n>/episode<n>.hdf5: head, left, and right camera RGB observations, head depth and segmentation, camera calibration, robot end pose, and joint actions.
raw/<task>/episode<n>/queries.npz and poses.npz: original point queries and simulator pose data used in label generation.
raw/<task>/episode<n>/instruction.json: task instructions.
labels_obj/<task>/episode<n>/queries.npz: eight object query points per window. anchor_xyz is [W,8,3]; future_delta_m is [W,8,32,3]. The file also includes image coordinates, entity/role IDs, validity and visibility masks, camera intrinsics, and frame IDs.
labels_ee/<task>/episode<n>/ee.npz: four kinematic points on each robot hand, ordered as base link 6, grasp midpoint, finger link 7, and finger link 8. current_xyz is [W,2,4,3]; future_delta_m is [W,2,4,32,3]. Validity masks are included.
windows.jsonl: one record per training window with relative paths and its episode, frame, window index, and split.
dataset_info.json: dataset counts and split rule.

Coordinates are in the current head-camera frame, in meters. Each future displacement is relative to that point's current XYZ anchor. Use the supplied validity masks when training or evaluating. Episode 0–48 of each task is the training split; episode 49 is reserved for diagnostics. Windows from the same episode share observations, so split by episode rather than by window.

The six tasks are adjust_bottle, handover_block, lift_pot, move_can_pot, open_laptop, and stack_blocks_two.

import json
import numpy as np

with open("windows.jsonl") as f:
    item = json.loads(next(f))
obj = np.load(item["object_labels"])
ee = np.load(item["ee_labels"])
w = item["window"]
object_start = obj["anchor_xyz"][w]
object_future_delta = obj["future_delta_m"][w]
ee_start = ee["current_xyz"][w]
ee_future_delta = ee["future_delta_m"][w]

Provenance

The episodes were collected with the RoboTwin platform; the object and end-effector trace labels were generated for the Trace-MoT project. See the RoboTwin project for platform details and attribution. Model checkpoints, DINO features, text embeddings, and OpenWAM training caches are outside the scope of this dataset.

Downloads last month: 59

Total file size:

4.35 GB

RoboTwin 300: object and end-effector point trajectories

Contents

Provenance