yujunjun/LLMs-from-scratch
1
0
Fork 0
mirror of https://github.com/rasbt/LLMs-from-scratch.git synced 2025-08-30 03:20:51 +00:00
Code Issues Packages Projects Releases Wiki Activity
LLMs-from-scratch/ch05
History
Daniel Kleine 7d0b9b78b0 Updated devcontainer, .gitignore and README for gutenberg project (#107) 
* added ch05/03_bonus_pretraining_on_gutenberg model checkpoints and preprocessing output folders to .gitignore

* removed prettier extension, added github alerts markdown extension

* specified download instructions and fixed code markdown

* Update ch05/03_bonus_pretraining_on_gutenberg/README.md

* Update ch05/03_bonus_pretraining_on_gutenberg/README.md

---------

Co-authored-by: Sebastian Raschka <mail@sebastianraschka.com>
2024-04-05 06:53:01 -05:00
..
01_main-chapter-code
Fix Loss in Gutenberg bonus section (#109)
2024-04-04 20:54:09 -05:00
02_alternative_weight_loading
Rename variable to context_length to make it easier on readers (#106)
2024-04-04 07:27:41 -05:00
03_bonus_pretraining_on_gutenberg
Updated devcontainer, .gitignore and README for gutenberg project (#107)
2024-04-05 06:53:01 -05:00
04_learning_rate_schedulers
Add and link bonus material (#84)
2024-03-23 07:27:43 -05:00
05_bonus_hparam_tuning
Rename variable to context_length to make it easier on readers (#106)
2024-04-04 07:27:41 -05:00
README.md
Add and link bonus material (#84)
2024-03-23 07:27:43 -05:00

README.md

Chapter 5: Pretraining on Unlabeled Data

  • 01_main-chapter-code contains the main chapter code
  • 02_alternative_weight_loading contains code to load the GPT model weights from alternative places in case the model weights become unavailable from OpenAI
  • 03_bonus_pretraining_on_gutenberg contains code to pretrain the LLM longer on the whole corpus of books from Project Gutenberg
  • [04_learning_rate_schedulers] contains code implementing a more sophisticated training function including learning rate schedulers and gradient clipping
  • 05_hparam_tuning contains an optional hyperparameter tuning script