Backup diff of BitNet (No. 5)

• Backup list
• View the diff current.
• View the diff for visual.
• View the source.
• View the backup.
• Go to BitNet.
  • 1 (2024-02-28 (Wed) 17:37:28)
  • 2 (2024-02-29 (Thu) 10:24:39)
  • 3 (2024-02-29 (Thu) 13:50:55)
  • 4 (2024-03-01 (Fri) 11:44:59)
  • 5 (2024-03-02 (Sat) 14:26:39)
  • 6 (2024-03-22 (Fri) 12:23:23)
  • 7 (2024-03-30 (Sat) 21:44:04)

• The added line is THIS COLOR.
• The deleted line is THIS COLOR.

TITLE:BitNet (Scaling 1-bit Transformers for Large Language Models)

- https://www.microsoft.com/en-us/research/publication/bitnet-scaling-1-bit-transformers-for-large-language-models/
- https://arxiv.org/abs/2310.11453
--  https://arxiv.org/pdf/2310.11453.pdf

- https://github.com/kyegomez/BitNet

*1.58 Bits [#t4c2ac80]

- https://arxiv.org/abs/2402.17764
--  https://arxiv.org/pdf/2402.17764.pdf

- https://github.com/Beomi/BitNet-Transformers/

- Microsoftが1.58ビットの大規模言語モデルをリリース、行列計算を足し算にできて計算コスト激減へ
--  https://gigazine.net/news/20240229-microsoft-1bit-llm/

- 1ビットLLMの衝撃! 70Bで8.9倍高速　全ての推論を加算のみで!GPU不要になる可能性も
--  https://topics.smt.docomo.ne.jp/article/wirelesswire/business/wirelesswire-20240286094?fm=latestnews


*. [#z7c55994]

- BitNet
--  https://qiita.com/fuyu_quant/items/fa4fc3930ddfd2ec6f16
>
※BitNetそのものを検証した記事ではなく，BitNetで使われているレイヤーを使った簡単なNNモデルを作成し検証しています．