Home/Skills/Preprocessing and Tokenizing Textual Datasets
tokenization

Preprocessing and Tokenizing Textual Datasets

Turning raw text into the tokenized, memory-mapped shards a GPT training loop can actually stream.

Free · download
Quick answer

Turning raw text into the tokenized, memory-mapped shards a GPT training loop can actually stream.

What this skill does

The preparation step that decides whether training is I/O bound before it starts: encoding, sharding, and the layout the loader reads.

Extracted by Loreto from github.com/karpathy/nanoGPT on 10 September 2026. One run, eighty six seconds, three skills — this is one of them. Nobody wrote it down first; it was read out of the source.

Free, with the whole package: SKILL.md, a README.md, and a generated test.

Get Preprocessing and Tokenizing Textual Datasets

Free to download — drops straight into .claude/skills/.