Pulse · rasbt/LLMs-from-scratch · GitHub

April 14, 2025 – April 21, 2025

Overview

6 Active pull requests

3 Active issues
- 6 Merged pull requests
- 0 Open pull requests
- 3 Closed issues
- 0 New issues

Could not load contribution data

Please try again later

6 Pull requests merged by 3 people

DeBERTa-v3 baseline
#630 merged Apr 20, 2025
BPE cosmetics
#629 merged Apr 18, 2025
Dpo vocab size clarification
#628 merged Apr 18, 2025
Llama3 from scratch improvements
#621 merged Apr 16, 2025
Minor DPO fixes
#617 merged Apr 16, 2025
fix: <|endoftext|> token
#620 merged Apr 16, 2025

3 Issues closed by 1 person

Byte-level BPE: Number of merges
#619 closed Apr 18, 2025
DPO: GPT-2 vocab size
#618 closed Apr 18, 2025
LayerNorm 'scale' and 'shift' attributes missing in GPTModel
#626 closed Apr 18, 2025