The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
Error code: DatasetGenerationError
Exception: ParserError
Message: Error tokenizing data. C error: Expected 2 fields in line 29, saw 23
Traceback: Traceback (most recent call last):
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/builder.py", line 1855, in _prepare_split_single
for _, table in generator:
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 690, in wrapped
for item in generator(*args, **kwargs):
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/packaged_modules/csv/csv.py", line 190, in _generate_tables
for batch_idx, df in enumerate(csv_file_reader):
File "/src/services/worker/.venv/lib/python3.9/site-packages/pandas/io/parsers/readers.py", line 1843, in __next__
return self.get_chunk()
File "/src/services/worker/.venv/lib/python3.9/site-packages/pandas/io/parsers/readers.py", line 1985, in get_chunk
return self.read(nrows=size)
File "/src/services/worker/.venv/lib/python3.9/site-packages/pandas/io/parsers/readers.py", line 1923, in read
) = self._engine.read( # type: ignore[attr-defined]
File "/src/services/worker/.venv/lib/python3.9/site-packages/pandas/io/parsers/c_parser_wrapper.py", line 234, in read
chunks = self._reader.read_low_memory(nrows)
File "parsers.pyx", line 850, in pandas._libs.parsers.TextReader.read_low_memory
File "parsers.pyx", line 905, in pandas._libs.parsers.TextReader._read_rows
File "parsers.pyx", line 874, in pandas._libs.parsers.TextReader._tokenize_rows
File "parsers.pyx", line 891, in pandas._libs.parsers.TextReader._check_tokenize_status
File "parsers.pyx", line 2061, in pandas._libs.parsers.raise_parser_error
pandas.errors.ParserError: Error tokenizing data. C error: Expected 2 fields in line 29, saw 23
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 1431, 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 992, in stream_convert_to_parquet
builder._prepare_split(
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/builder.py", line 1742, in _prepare_split
for job_id, done, content in self._prepare_split_single(
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/builder.py", line 1898, 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.
ep_rew_mean
float64 | total_timesteps
int64 |
|---|---|
-36.7 | 6,400 |
-39.1 | 12,800 |
-37.7 | 19,200 |
-36.9 | 25,600 |
-33.7 | 32,000 |
-30.5 | 38,400 |
-28.3 | 44,800 |
-25.8 | 51,200 |
-25 | 57,600 |
-23.8 | 64,000 |
-22.5 | 70,400 |
-20.8 | 76,800 |
-19.5 | 83,200 |
-18.5 | 89,600 |
-17.8 | 96,000 |
-16.6 | 102,400 |
-14 | 108,800 |
-8.35 | 115,200 |
-8 | 121,600 |
-2.01 | 128,000 |
-7.58 | 134,400 |
0.267 | 140,800 |
-1.33 | 147,200 |
10.3 | 153,600 |
9.98 | 160,000 |
48.7 | 166,400 |
63.5 | 172,800 |
108 | 179,200 |
80.2 | 185,600 |
405 | 192,000 |
1,060 | 198,400 |
1,530 | 204,800 |
1,620 | 211,200 |
3,340 | 217,600 |
4,550 | 224,000 |
7,370 | 230,400 |
9,150 | 236,800 |
12,100 | 243,200 |
15,900 | 249,600 |
17,800 | 256,000 |
23,200 | 262,400 |
25,800 | 268,800 |
28,200 | 275,200 |
28,300 | 281,600 |
30,000 | 288,000 |
33,800 | 294,400 |
37,000 | 300,800 |
38,400 | 307,200 |
38,800 | 313,600 |
37,800 | 320,000 |
36,400 | 326,400 |
37,000 | 332,800 |
36,700 | 339,200 |
38,300 | 345,600 |
37,900 | 352,000 |
37,500 | 358,400 |
39,800 | 364,800 |
42,300 | 371,200 |
45,800 | 377,600 |
46,300 | 384,000 |
47,400 | 390,400 |
48,000 | 396,800 |
50,800 | 403,200 |
53,400 | 409,600 |
54,700 | 416,000 |
53,100 | 422,400 |
53,300 | 428,800 |
54,000 | 435,200 |
52,500 | 441,600 |
52,900 | 448,000 |
51,900 | 454,400 |
51,800 | 460,800 |
51,700 | 467,200 |
51,600 | 473,600 |
52,300 | 480,000 |
51,900 | 486,400 |
52,500 | 492,800 |
54,400 | 499,200 |
54,600 | 505,600 |
53,300 | 512,000 |
53,500 | 518,400 |
52,900 | 524,800 |
54,100 | 531,200 |
53,400 | 537,600 |
54,200 | 544,000 |
54,800 | 550,400 |
54,800 | 556,800 |
53,700 | 563,200 |
52,800 | 569,600 |
54,100 | 576,000 |
56,100 | 582,400 |
58,300 | 588,800 |
61,700 | 595,200 |
64,000 | 601,600 |
66,800 | 608,000 |
67,500 | 614,400 |
71,000 | 620,800 |
72,300 | 627,200 |
71,200 | 633,600 |
70,700 | 640,000 |
🧩 Introduction
This is the code base for our paper on [Learning Fused State Representations for Control from Multi-View Observations]. we propose Multi-view Fusion State for Control (MFSC), firstly incorporating bisimulation metrics learning into MVRL to learn task-relevant representations. Furthermore, we propose a multiview-based mask and latent reconstruction auxiliary task that exploits shared information across views and improves MFSC’s robustness in missing views by introducing a mask token.
⚒️ Installation
For installation and setup instructions for each environment, please refer to the corresponding subdirectories under envs/ or their README files. You may also refer to the setup guidelines from Keypoint3D and DBC for additional reference and compatibility.
📖 Run Experiments
We evaluate our method on a set of 3D manipulation environments Meta-World, a high degree of freedom 3D locomotion environment PyBullet's Ant, and a more realistic multi-view highway driving scenario, CARLA.
To train MFSC from scratch on each benchmark, simply execute the corresponding .sh script located in its respective directory:
# Meta-world
$ bash run.sh
# Pybullet's Ant
$ bash run.sh
# CARLA
bash run_local_carla096.sh
🚀 Checkpoints and Original Data
We have made all original training log data, along with intermediate model checkpoints, available in our Hugging Face repository. We hope this resource is helpful for your experiments and further research.
📌 Citation
If you find this work useful for your research, please consider citing it.
@article{wang2025learning,
title={Learning Fused State Representations for Control from Multi-View Observations},
author={Wang, Zeyu and Li, Yao-Hui and Li, Xin and Zang, Hongyu and Laroche, Romain and Islam, Riashat},
journal={arXiv preprint arXiv:2502.01316},
year={2025}
}
👍 Acknowledgments
Thanks to Keypoint3D, DBC, SimSR and MLR for their great work and codebase, which served as the foundation for developing MFSC.
📧 Contact Us
If you have any question, please feel free to contact us via [email protected].
- Downloads last month
- 204