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
t: int64
loss: double
td_abs: double
td_clipped: double
q_taken: double
q_max: double
target: double
dead_relu: double
grad_norm: double
clipped: double
update_ratio: double
param_norm: double
updates: int64
evaluations: null
probes: null
train: null
food_curriculum: null
distractors: int64
seed: int64
actions: string
diagnostics: list<item: null>
child 0, item: null
steps: int64
wall_scale: double
eval: struct<every: int64, episodes: int64, eps: double>
child 0, every: int64
child 1, episodes: int64
child 2, eps: double
opt: string
bonus: double
eps_decay: null
food_relocate: bool
lr: null
eps_floor: double
frame_stack: null
probe_far: bool
agent: string
size: int64
render: string
map: string
episodes: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
wallclock_s: double
to
{'agent': Value('string'), 'seed': Value('int64'), 'steps': Value('int64'), 'distractors': Value('int64'), 'size': Value('int64'), 'eps_floor': Value('float64'), 'eps_decay': Value('null'), 'food_curriculum': Value('null'), 'food_relocate': Value('bool'), 'map': Value('string'), 'bonus': Value('float64'), 'render': Value('string'), 'frame_stack': Value('null'), 'wall_scale': Value('float64'), 'actions': Value('string'), 'probe_far': Value('bool'), 'opt': Value('string'), 'lr': Value('null'), 'episodes': List(List(Value('int64'))), 'diagnostics': List(Value('null')), 'eval': {'every': Value('int64'), 'episodes': Value('int64'), 'eps': Value('float64')}, 'evaluations': List(List(Json(decode=True))), 'probes': List(List(Json(decode=True))), 'train': List(List(Json(decode=True))), 'wallclock_s': Value('float64')}
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
t: int64
loss: double
td_abs: double
td_clipped: double
q_taken: double
q_max: double
target: double
dead_relu: double
grad_norm: double
clipped: double
update_ratio: double
param_norm: double
updates: int64
evaluations: null
probes: null
train: null
food_curriculum: null
distractors: int64
seed: int64
actions: string
diagnostics: list<item: null>
child 0, item: null
steps: int64
wall_scale: double
eval: struct<every: int64, episodes: int64, eps: double>
child 0, every: int64
child 1, episodes: int64
child 2, eps: double
opt: string
bonus: double
eps_decay: null
food_relocate: bool
lr: null
eps_floor: double
frame_stack: null
probe_far: bool
agent: string
size: int64
render: string
map: string
episodes: list<item: list<item: int64>>
child 0, item: list<item: int64>
child 0, item: int64
wallclock_s: double
to
{'agent': Value('string'), 'seed': Value('int64'), 'steps': Value('int64'), 'distractors': Value('int64'), 'size': Value('int64'), 'eps_floor': Value('float64'), 'eps_decay': Value('null'), 'food_curriculum': Value('null'), 'food_relocate': Value('bool'), 'map': Value('string'), 'bonus': Value('float64'), 'render': Value('string'), 'frame_stack': Value('null'), 'wall_scale': Value('float64'), 'actions': Value('string'), 'probe_far': Value('bool'), 'opt': Value('string'), 'lr': Value('null'), 'episodes': List(List(Value('int64'))), 'diagnostics': List(Value('null')), 'eval': {'every': Value('int64'), 'episodes': Value('int64'), 'eps': Value('float64')}, 'evaluations': List(List(Json(decode=True))), 'probes': List(List(Json(decode=True))), 'train': List(List(Json(decode=True))), 'wallclock_s': Value('float64')}
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.
Snake RL results
Raw outputs of the study project felipedreis/snake_rl: Neural Episodic Control
(Pritzel et al., 2017), DQN and MFEC on a small Snake grid, in pure NumPy. The layout mirrors the repo's results/
directory: results/<board>/<agent>_s<seed>.json (episodes, diagnostics, probes, training telemetry), *.train.jsonl
(telemetry streamed during a run), and *.agent.pkl (trained agents, without their replay buffers).
To use it with the repo: hf download felipedreis/snake-rl-results --repo-type dataset --local-dir results, then
snake-plot results/d0 or snake-watch results/d0/dqn_cnn_s1.agent.pkl.
Warning about the .agent.pkl files
They are Python pickles. Loading a pickle runs arbitrary code, so only load these files if you trust this repository, and never load pickles from sources you do not trust. The JSON and JSONL files are plain data.
- Downloads last month
- 2,103