---
dataset_info:
  features:
  - name: input
    dtype: string
  - name: edits
    sequence: string
  - name: output
    dtype: string
  splits:
  - name: train
    num_bytes: 5685159
    num_examples: 509
  download_size: 1256858
  dataset_size: 5685159
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: apache-2.0
language:
- en
tags:
- code
size_categories:
- n<1K
---

A synthetic dataset of inline edit predictions based on [kalinkov/tailwindcss_components](https://huggingface.co/datasets/kalinkov/tailwindcss_components) inspired by the strategy of [Zed's model](https://zed.dev/blog/edit-prediction)