Hugging Face's logo

Datasets:
DarthVaderSenior
/
visgym-pacman-2d

Tasks:
Image-to-Text
Reinforcement Learning
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Preview
Duplicate
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
task: string
env_id: string
env_kwargs: struct<food_count: int64, max_steps: int64>
  child 0, food_count: int64
  child 1, max_steps: int64
seed: int64
episode_seed: int64
extra_state: null
timestamp: string
history: list<item: struct<step: int64, prompt: string, vlm_output: string, action: string, reward: double, i (... 68 chars omitted)
  child 0, item: struct<step: int64, prompt: string, vlm_output: string, action: string, reward: double, image_prev:  (... 56 chars omitted)
      child 0, step: int64
      child 1, prompt: string
      child 2, vlm_output: string
      child 3, action: string
      child 4, reward: double
      child 5, image_prev: string
      child 6, image: string
      child 7, image_next: string
      child 8, info: string
stats: struct<step: int64, reward: double, terminated: bool, truncated: bool>
  child 0, step: int64
  child 1, reward: double
  child 2, terminated: bool
  child 3, truncated: bool
init_state: struct<backend: string, difficulty: string, grid_size: int64, food_count: int64, wall_density: doubl (... 265 chars omitted)
  child 0, backend: string
  child 1, difficulty: string
  child 2, grid_size: int64
  child 3, food_count: int64
  child 4, wall_density: double
  child 5, max_steps: int64
  child 6, horizon: int64
  child 7, tile_size: int64
  child 8, walls: list<item: list<item: int64>>
      child 0, item: list<item: int64>
          child 0, item: int64
  child 9, food: list<item: list<item: int64>>
      child 0, item: list<item: int64>
          child 0, item: int64
  child 10, pacman_pos: list<item: int64>
      child 0, item: int64
  child 11, ghost_pos: list<item: int64>
      child 0, item: int64
  child 12, pacman_dir: int64
  child 13, ghost_dir: int64
  child 14, steps: int64
  child 15, alive: bool
  child 16, won: bool
solve_kwargs: struct<strategy: string, horizon: int64>
  child 0, strategy: string
  child 1, horizon: int64
version: string
episode: int64
difficulty: string
dataset_pipeline: string
init_state_hash: string
solve_info: struct<strategy_requested: string, planner_used: string, fallback_planner: string, horizon: int64, m (... 36 chars omitted)
  child 0, strategy_requested: string
  child 1, planner_used: string
  child 2, fallback_planner: string
  child 3, horizon: int64
  child 4, max_moves: int64
  child 5, plan_length: int64
generated_at: string
source_split: string
source_record_index: int64
source_hash: string
to
{'task': Value('string'), 'env_id': Value('string'), 'env_kwargs': {'food_count': Value('int64'), 'max_steps': Value('int64')}, 'seed': Value('int64'), 'episode_seed': Value('int64'), 'extra_state': Value('null'), 'timestamp': Value('string'), 'history': List({'step': Value('int64'), 'prompt': Value('string'), 'vlm_output': Value('string'), 'action': Value('string'), 'reward': Value('float64'), 'image_prev': Value('string'), 'image': Value('string'), 'image_next': Value('string'), 'info': Value('string')}), 'stats': {'step': Value('int64'), 'reward': Value('float64'), 'terminated': Value('bool'), 'truncated': Value('bool')}, 'init_state': {'backend': Value('string'), 'difficulty': Value('string'), 'grid_size': Value('int64'), 'food_count': Value('int64'), 'wall_density': Value('float64'), 'max_steps': Value('int64'), 'horizon': Value('int64'), 'tile_size': Value('int64'), 'walls': List(List(Value('int64'))), 'food': List(List(Value('int64'))), 'pacman_pos': List(Value('int64')), 'ghost_pos': List(Value('int64')), 'pacman_dir': Value('int64'), 'ghost_dir': Value('int64'), 'steps': Value('int64'), 'alive': Value('bool'), 'won': Value('bool')}, 'solve_kwargs': {'strategy': Value('string'), 'horizon': Value('int64')}, 'version': Value('string'), 'episode': Value('int64'), 'difficulty': Value('string'), 'dataset_pipeline': Value('string'), 'init_state_hash': Value('string'), 'solve_info': {'strategy_requested': Value('string'), 'planner_used': Value('string'), 'fallback_planner': Value('string'), 'horizon': Value('int64'), 'max_moves': Value('int64'), 'plan_length': Value('int64')}, 'generated_at': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1779, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 609, in wrapped
                  for item in generator(*args, **kwargs):
                              ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 295, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 128, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2321, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2249, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              task: string
              env_id: string
              env_kwargs: struct<food_count: int64, max_steps: int64>
                child 0, food_count: int64
                child 1, max_steps: int64
              seed: int64
              episode_seed: int64
              extra_state: null
              timestamp: string
              history: list<item: struct<step: int64, prompt: string, vlm_output: string, action: string, reward: double, i (... 68 chars omitted)
                child 0, item: struct<step: int64, prompt: string, vlm_output: string, action: string, reward: double, image_prev:  (... 56 chars omitted)
                    child 0, step: int64
                    child 1, prompt: string
                    child 2, vlm_output: string
                    child 3, action: string
                    child 4, reward: double
                    child 5, image_prev: string
                    child 6, image: string
                    child 7, image_next: string
                    child 8, info: string
              stats: struct<step: int64, reward: double, terminated: bool, truncated: bool>
                child 0, step: int64
                child 1, reward: double
                child 2, terminated: bool
                child 3, truncated: bool
              init_state: struct<backend: string, difficulty: string, grid_size: int64, food_count: int64, wall_density: doubl (... 265 chars omitted)
                child 0, backend: string
                child 1, difficulty: string
                child 2, grid_size: int64
                child 3, food_count: int64
                child 4, wall_density: double
                child 5, max_steps: int64
                child 6, horizon: int64
                child 7, tile_size: int64
                child 8, walls: list<item: list<item: int64>>
                    child 0, item: list<item: int64>
                        child 0, item: int64
                child 9, food: list<item: list<item: int64>>
                    child 0, item: list<item: int64>
                        child 0, item: int64
                child 10, pacman_pos: list<item: int64>
                    child 0, item: int64
                child 11, ghost_pos: list<item: int64>
                    child 0, item: int64
                child 12, pacman_dir: int64
                child 13, ghost_dir: int64
                child 14, steps: int64
                child 15, alive: bool
                child 16, won: bool
              solve_kwargs: struct<strategy: string, horizon: int64>
                child 0, strategy: string
                child 1, horizon: int64
              version: string
              episode: int64
              difficulty: string
              dataset_pipeline: string
              init_state_hash: string
              solve_info: struct<strategy_requested: string, planner_used: string, fallback_planner: string, horizon: int64, m (... 36 chars omitted)
                child 0, strategy_requested: string
                child 1, planner_used: string
                child 2, fallback_planner: string
                child 3, horizon: int64
                child 4, max_moves: int64
                child 5, plan_length: int64
              generated_at: string
              source_split: string
              source_record_index: int64
              source_hash: string
              to
              {'task': Value('string'), 'env_id': Value('string'), 'env_kwargs': {'food_count': Value('int64'), 'max_steps': Value('int64')}, 'seed': Value('int64'), 'episode_seed': Value('int64'), 'extra_state': Value('null'), 'timestamp': Value('string'), 'history': List({'step': Value('int64'), 'prompt': Value('string'), 'vlm_output': Value('string'), 'action': Value('string'), 'reward': Value('float64'), 'image_prev': Value('string'), 'image': Value('string'), 'image_next': Value('string'), 'info': Value('string')}), 'stats': {'step': Value('int64'), 'reward': Value('float64'), 'terminated': Value('bool'), 'truncated': Value('bool')}, 'init_state': {'backend': Value('string'), 'difficulty': Value('string'), 'grid_size': Value('int64'), 'food_count': Value('int64'), 'wall_density': Value('float64'), 'max_steps': Value('int64'), 'horizon': Value('int64'), 'tile_size': Value('int64'), 'walls': List(List(Value('int64'))), 'food': List(List(Value('int64'))), 'pacman_pos': List(Value('int64')), 'ghost_pos': List(Value('int64')), 'pacman_dir': Value('int64'), 'ghost_dir': Value('int64'), 'steps': Value('int64'), 'alive': Value('bool'), 'won': Value('bool')}, 'solve_kwargs': {'strategy': Value('string'), 'horizon': Value('int64')}, 'version': Value('string'), 'episode': Value('int64'), 'difficulty': Value('string'), 'dataset_pipeline': Value('string'), 'init_state_hash': Value('string'), 'solve_info': {'strategy_requested': Value('string'), 'planner_used': Value('string'), 'fallback_planner': Value('string'), 'horizon': Value('int64'), 'max_moves': Value('int64'), 'plan_length': Value('int64')}, 'generated_at': 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 1342, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 907, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1646, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1832, 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.

task
string
env_id
string
env_kwargs
dict
seed
int64
episode_seed
int64
extra_state
null
timestamp
string
history
list
stats
dict
init_state
dict
solve_kwargs
dict
version
string
episode
int64
difficulty
string
dataset_pipeline
string
init_state_hash
string
solve_info
dict
generated_at
string
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,000
null
2026-04-25T06:33:02.488128
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 23, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
0
ultrahard
behavior_cloning
48e6461420ededb9ab4a46d4784fdb3ee847c7e63fb9874fcd44ba43bcbac7c3
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:02.489028
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,001
null
2026-04-25T06:33:02.817067
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 10, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
1
ultrahard
behavior_cloning
e742dcc47146f9d8a30cf195f85290eec5235b0a262347eb2839d779eec8cf43
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:02.817730
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,002
null
2026-04-25T06:33:05.756044
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 23, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
2
ultrahard
behavior_cloning
020f6c0c2ec99307a2e74349533044189376ade762e35cb1060a697df6483a64
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:05.756731
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,003
null
2026-04-25T06:33:07.010168
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 18, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
3
ultrahard
behavior_cloning
80089144f31cb13acb0d45b42893e89edc9de97987a2b28334815a9857464a08
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:07.010890
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,004
null
2026-04-25T06:33:08.250314
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 17, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
4
ultrahard
behavior_cloning
f9301c064d9d4325749713d6eed81725cac11447a4949c048cc9630a84cbfb30
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:08.250958
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,005
null
2026-04-25T06:33:10.886047
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 27, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
5
ultrahard
behavior_cloning
eb04d6e89f3f34ea42435d1c943f16e5162c823ba0b6b9511f15db626f22e5f3
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:10.886514
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,006
null
2026-04-25T06:33:11.070021
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 10, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
6
ultrahard
behavior_cloning
1efcb7097da4501a7f10f4b552b6528db84eed76ccc9f904d9f70765bfd2f39e
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:11.070645
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,007
null
2026-04-25T06:33:14.101787
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 16, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
7
ultrahard
behavior_cloning
a02dbeb50ccae4b4bdc783929d27bf37294eafbe909fe966f609d494ed763e1f
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:14.102429
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,008
null
2026-04-25T06:33:14.655343
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 12, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
8
ultrahard
behavior_cloning
410f8f087158cc42e97663e73d99d6818aff18581c19a62f6f485eb463103c9f
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:14.655987
pacman_2d_ultrahard_v0
pacman_2d/hard
{ "food_count": 8, "max_steps": 40 }
20,000,000
20,000,009
null
2026-04-25T06:33:15.016717
[{"step":0,"prompt":"You are Pacman in a deterministic 11x11 maze. Collect every food pellet, avoid (...TRUNCATED)
{ "step": 12, "reward": 1, "terminated": true, "truncated": false }
{"backend":"custom_grid_pacman_python_assets","difficulty":"hard","grid_size":11,"food_count":8,"wal(...TRUNCATED)
{ "strategy": "receding_horizon", "horizon": 8 }
visgym_repo_solver_v1
9
ultrahard
behavior_cloning
cb3436ab3bc806846ff408d2e7540cdf1bd4c67a94d2370f589b55ebb504fec8
{"strategy_requested":"receding_horizon","planner_used":"shortest_safe","fallback_planner":"receding(...TRUNCATED)
2026-04-25T06:33:15.017325
End of preview.

VisGym Pacman2D Deterministic Trajectories

This dataset contains behavior-cloning trajectories for the custom VisGym Pacman2D environment.

Each row is one successful oracle episode. The history entries include image_prev, image, and image_next; no synthetic thinking traces are stored. Images are JPEG base64 strings rendered from the greyblue9/pacman-python visual assets used by the environment.

Environment summary:

  • Grid size is constrained to 7x7 through 12x12.
  • Easy uses the 9x9 setting.
  • Hard uses the 11x11 setting with 8 food pellets.
  • Ultrahard uses the 11x11 hard layout with 8 food pellets and a 40-step budget.
  • Entities are Pacman, sparse food, walls, and exactly one deterministic ghost.
  • The ghost chases Pacman by BFS distance with fixed tie-break order up, down, left, right.
  • The movement action space is four directions; VisGym records can also end with ('stop', 'stop').

Splits:

  • data/pacman_2d_easy_v0/train/*.jsonl.gz
  • data/pacman_2d_easy_v0/test/*.jsonl.gz
  • data/pacman_2d_hard_v0/train/*.jsonl.gz
  • data/pacman_2d_hard_v0/test/*.jsonl.gz
  • data/pacman_2d_ultrahard_v0/train/*.jsonl.gz
  • data/pacman_2d_ultrahard_v0/test/*.jsonl.gz

Train/test initial states are checked to be disjoint by init_state_hash; see metadata/init_state_hash_audit.json.

Target Hub repo: https://huggingface.co/datasets/DarthVaderSenior/visgym-pacman-2d

Final Test Split

A deterministic final_test split was added on 2026-05-04T20:18:20Z for fixed checkpoint evaluation.

Split Task Records Source split Selection policy Manifest
final_test pacman_2d_ultrahard_v0 100 test First 100 records in stable shard order metadata/final_test_selection.json

The final-test records preserve the original trajectory and image fields. Top-level provenance fields identify the source split, source record index, and original record SHA-256 hash.
Downloads last month
1,574
Total file size:
158 GB