Code

Implement gemma2 training#127

Code

Open

bzantium wants to merge 7 commits intoMcGill-NLP:mainMcGill-NLP/llm2vec:mainfrom bzantium:feature/#126bzantium/llm2vec:feature/#126Copy head branch name to clipboard

+217-29

Commits

Commits on Jul 18, 2024

• implement gemma2 training

  

  ryan.agile

  committed
  f4f0a9d

  Copy full SHA for f4f0a9d

• implement gemma2 training

  

  ryan.agile

  committed
  ee621c1

  Copy full SHA for ee621c1

• implement gemma2 training

  

  ryan.agile

  committed
  d7fb636

  Copy full SHA for d7fb636

• assign config attribute and remove unnecessary part

  

  ryan.agile

  committed
  f4a8fce

  Copy full SHA for f4a8fce

• revise ModifiedGemma2DecoderLayer

  

  ryan.agile

  committed
  bbcf43f

  Copy full SHA for bbcf43f

Commits on Jul 19, 2024

• fix _update_causal_mask for gemma modeling

  

  ryan.agile

  committed
  c62c2f0

  Copy full SHA for c62c2f0

Commits on Jul 24, 2024

• revise llama modeling code for Llama3.1

  

  ryan.agile

  committed
  03382c3

  Copy full SHA for 03382c3