---
configs:
- config_name: train
  default: true
  data_files:
  - split: train
    path: train.jsonl

- config_name: math_eval
  data_files:
  - split: test
    path: math_eval.jsonl

- config_name: gsm8k_eval
  data_files:
  - split: test
    path: gsm8k_eval.jsonl
---

# Your dataset name

Combined GSM8K + MATH training data, with separate eval configs.
```
from datasets import load_dataset

# From Hub
train = load_dataset("Ritvik19/gsm8k-math", "train", split="train")
math_eval = load_dataset("Ritvik19/gsm8k-math", "math_eval", split="test")
gsm8k_eval = load_dataset("Ritvik19/gsm8k-math", "gsm8k_eval", split="test")
```