Update README.md

README.md

@@ -52,3 +52,52 @@ configs:
   - split: test
     path: data/test-*
 ---
+
+# Derivative Dataset
+
+## Dataset Summary
+
+The Derivative dataset is a collection of functions and their derivatives. The dataset
+is generated using Gaussian random functions and their derivatives.
+
+Plot from the dataset:
+
+![Derivative Dataset](plot.png)
+
+## Dataset Structure
+
+The dataset is structured as follows.
+
+### Subsets
+
+The dataset is divided into two subsets:
+
+- `default`: the default dataset.
+- `aligned`: the `X` values are aligned for all functions.
+
+### Splits
+
+The dataset is divided into two splits:
+
+- `train`: the training split, consisting of 10,000 rows.
+- `test`: the test split, consisting of 1,000 rows.
+
+### Columns
+
+Each row in the dataset consists of the following columns:
+
+- `X`: the input value of the function f.
+- `f`: the value of the function f at X.
+- `Y`: the input value of the derivative f'.
+- `f'`: the value of the derivative f' at Y.
+
+### Data
+
+The dataset contains the following data:
+
+| Column  | Shape   | Type    |
+|---------|---------|---------|
+| X       | (100,)  | float   |
+| f       | (100,)  | float   |
+| Y       | (100,)  | float   |
+| f'      | (100,)  | float   |