mirror of https://github.com/rasbt/LLMs-from-scratch.git synced 2026-04-10 12:33:42 +00:00

Files

Maxwell De Jong e0dbec3331 Fix encoding of multiple preceding spaces in BPE tokenizer. (#945 )

* Fix encoding of multiple preceding spaces in BPE tokenizer.

* Add test

---------

Co-authored-by: rasbt <mail@sebastianraschka.com>

2026-01-10 10:27:23 -06:00

2025-10-07 15:22:59 -05:00

2025-10-21 19:42:33 -05:00

minor spelling fix

2024-09-08 15:35:36 -05:00

fixed num_workers (#229 )

2024-06-19 17:36:46 -05:00

2026-01-10 10:27:23 -06:00

README.md

fixed video link (#646 )

2025-06-13 08:16:18 -05:00

Chapter 2: Working with Text Data

Main Chapter Code

02_bonus_bytepair-encoder contains optional code to benchmark different byte pair encoder implementations
03_bonus_embedding-vs-matmul contains optional (bonus) code to explain that embedding layers and fully connected layers applied to one-hot encoded vectors are equivalent.
04_bonus_dataloader-intuition contains optional (bonus) code to explain the data loader more intuitively with simple numbers rather than text.
05_bpe-from-scratch contains (bonus) code that implements and trains a GPT-2 BPE tokenizer from scratch.

In the video below, I provide a code-along session that covers some of the chapter contents as supplementary material.