Code

Fixing block size for Mistral-7B.#141

Code

Open

Artyom17 wants to merge 3 commits intopytorch-labs:mainpytorch-labs/gpt-fast:mainfrom SesameAILabs:art/fix-mistralSesameAILabs/gpt-fast:art/fix-mistralCopy head branch name to clipboard

+2-2

Commits

Commits on Mar 19, 2024

• Fixing block size for Mistral-7B.

  Artyom17

  committed
  cc4cbec

  Copy full SHA for cc4cbec

• Saving some memory on freq_cis tensor with big block_size but small max_seq_length

  Artyom17

  committed
  8ca548c

  Copy full SHA for 8ca548c

Commits on Apr 4, 2024

• Merge branch 'main' into art/fix-mistral

  Artyom17

  authored
  f21da73

  Copy full SHA for f21da73