# longhealth-longcode-19k-max-len55k-enc-max-len1k-dec

max len55k enc max len1k dec

## Dataset Structure

This dataset is split into multiple parts due to size constraints:
- Training data: `train_part*.jsonl`
- Validation data: `val_part*.jsonl`

Each part is approximately 3GB or smaller.