Upload README.md with huggingface_hub

README.md

@@ -1,3 +1,92 @@
----
-license: mit
----
+---
+pretty_name: Permutation Groups Benchmark
+tags:
+  - mathematics
+  - group-theory
+  - permutations
+  - sequence-to-sequence
+  - benchmark
+  - generated
+task_categories:
+  - text-generation
+  - sequence-modeling
+licenses:
+  - mit
+configs:
+  - config_name: s3_data
+    data_files:
+      train: s3_data/train-*.parquet
+      test: s3_data/test-*.parquet
+    description: Permutation Composition Dataset for the Symmetric Group S3.
+  - config_name: s4_data
+    data_files:
+      train: s4_data/train-*.parquet
+      test: s4_data/test-*.parquet
+    description: Permutation Composition Dataset for the Symmetric Group S4.
+  - config_name: s5_data
+    data_files:
+      train: s5_data/train-*.parquet
+      test: s5_data/test-*.parquet
+    description: Permutation Composition Dataset for the Symmetric Group S5.
+  - config_name: s6_data
+    data_files:
+      train: s6_data/train-*.parquet
+      test: s6_data/test-*.parquet
+    description: Permutation Composition Dataset for the Symmetric Group S6.
+  - config_name: s7_data
+    data_files:
+      train: s7_data/train-*.parquet
+      test: s7_data/test-*.parquet
+    description: Permutation Composition Dataset for the Symmetric Group S7.
+  - config_name: a5_data
+    data_files:
+      train: a5_data/train-*.parquet
+      test: a5_data/test-*.parquet
+    description: Permutation Composition Dataset for the Alternating Group A5.
+  - config_name: a6_data
+    data_files:
+      train: a6_data/train-*.parquet
+      test: a6_data/test-*.parquet
+    description: Permutation Composition Dataset for the Alternating Group A6.
+  - config_name: a7_data
+    data_files:
+      train: a7_data/train-*.parquet
+      test: a7_data/test-*.parquet
+    description: Permutation Composition Dataset for the Alternating Group A7.
+--- 
+
+# Permutation Groups Benchmark Dataset
+
+This repository contains a collection of permutation composition datasets for various symmetric and alternating groups. These datasets are designed for benchmarking sequence-to-sequence models on group theory tasks.
+
+## Dataset Configurations
+
+Each group (e.g., S3, A5) is available as a separate configuration within this dataset. You can load a specific configuration using the `name` argument in `load_dataset`:
+
+```python
+from datasets import load_dataset
+
+# Load the S3 dataset
+s3_dataset = load_dataset("BeeGass/permutation-groups", name="s3_data")
+
+# Load the A5 dataset
+a5_dataset = load_dataset("BeeGass/permutation-groups", name="a5_data")
+
+# Load all datasets (as a DatasetDict of DatasetDicts)
+all_datasets = load_dataset("BeeGass/permutation-groups")
+```
+
+## Data Structure
+
+Each dataset configuration contains `train` and `test` splits. Each sample in the dataset has the following features:
+
+- `input_sequence`: A space-separated string of integer IDs representing the sequence of permutations to be composed.
+- `target`: An integer ID representing the composition of the `input_sequence` permutations.
+
+## Permutation Mapping
+
+For each configuration, the mapping from integer IDs to their corresponding permutation array forms is provided in a `metadata.json` file located within the respective subfolder (e.g., `s3_data/metadata.json`). This file is crucial for interpreting the `input_sequence` and `target` IDs.
+
+## License
+
+This dataset is licensed under the MIT License.