Skip to content
ToolQivo

Dataset Splitter

Split a dataset into training, validation and test files with a repeatable shuffle.

AI Tools Reviews Free · no sign-up

Or drop a file — it is split in your browser, never uploaded

Test gets whatever is left. The same seed always produces the same split.

Common splits

Quick Tool Overview

The Dataset Splitter divides JSONL, CSV or text data into training, validation and test files with a seeded shuffle, so the split is repeatable, removes duplicates first and repeats CSV headers in every file.

What it does
Splits a dataset into train, validation and test
You give it
A file or pasted rows and the percentages
You get back
Three files, previewed and downloadable
Best for
Machine learning, fine-tuning and evaluation sets
AI Tools Free, unlimited use No sign-up Runs in your browser

About the Dataset Splitter

Before training or evaluating a model you need data it has never seen: a validation set to tune against and a test set to report on. Splitting by hand invites the classic mistakes — the file was sorted, so validation looks nothing like training, or duplicates land on both sides and inflate every score.

Paste or drop a JSONL, CSV or plain text file with one example per line, choose the percentages, and get three files. The rows are shuffled with a seeded generator, so the same seed always produces exactly the same split — which matters when you rerun an experiment and need the test set to stay put. Exact duplicates can be removed before splitting, and a CSV header line can be repeated at the top of every file.

Rounding never drops rows: whatever is left over goes to training, so the three parts always add back up to your total. Each part can be previewed, copied or downloaded with its original file extension.

Key features of the Dataset Splitter

Repeatable seeded shuffle

The same seed always produces the same split, so test sets stay stable across runs.

No rows lost to rounding

The remainder goes to training, so the three parts always add up to the total.

Duplicate removal

Exact duplicates are removed first, so no example sits on both sides of the split.

CSV header support

The header line is repeated at the top of each output file.

Common presets

80/10/10, 70/15/15, 90/10/0 and 80/20/0 in one click.

How to use the Dataset Splitter

  1. 1

    Add the dataset

    Paste one example per line, or drop a .jsonl, .csv or .txt file.

  2. 2

    Choose the percentages

    Set train and validation; test gets the remainder. Presets cover common splits.

  3. 3

    Set the seed

    Keep the same seed to reproduce the split exactly later.

  4. 4

    Download the three files

    Each part downloads separately, named after your original file.

Frequently asked questions

80/10/10 is the common default. With small datasets, 70/15/15 gives evaluation sets large enough to mean something; with very large datasets, a smaller share such as 98/1/1 is often enough.

Reviews of the Dataset Splitter

What people who use the Dataset Splitter think of it. Reviews come from visitors and are posted as written.

No ratings yet

Yours would be the first.

  • 5 0 reviews rated 5 stars
  • 4 0 reviews rated 4 stars
  • 3 0 reviews rated 3 stars
  • 2 0 reviews rated 2 stars
  • 1 0 reviews rated 1 star

Write a review

Used this tool? Your rating helps the next person decide.

Your rating

Plain text only — links, email addresses and HTML are not accepted. Your name and review are shown publicly; nothing else is collected. See the privacy policy.

0 reviews

Newest first

No reviews yet

Be the first to rate the Dataset Splitter. It takes half a minute and no account.

Write the first review

More AI tools