Rare find

GLM. An AI model that learns by filling in blanks and works across writing and reading tasks.

github.com/THUDM/GLM

Vaya's read on this project

Problem, audience, market, and the verdict — sign in to see it.

Updates

March 2023
  • Update README
  • Delete print
  • Fix attention_mask for seq2seq tasks when mask_pad_token
  • Fix model parallelism in finetuning
  • Add deepspeed.init_distributed
  • Add version for DeepSpeed
  • Update README.md
  • Update README.md
February 2023
  • Skip loading random state in generate_samples.py
January 2023
  • Add top_p and top_k for seq2seq evaluation
  • Add mask-pad-token for seq2seq tasks
  • Add mask-pad-token for seq2seq tasks
  • Update model log
  • Add conversion script
  • Update OneDrive link
  • Add padding option
  • Fix corpus for glm-large-chinese in README.md
December 2022
  • Add default label for multichoice dataset
  • Merge pull request #48 from TrellixVulnTeam/main
  • Add script for converting checkpoint