transformers
add VisionTextDualEncoder and CLIP fine-tuning script
#15701
Merged
Overview
Commits
21
Changes
View On GitHub
Commits
  • begin script
    patil-suraj committed 4 years ago
  • update script
    patil-suraj committed 4 years ago
  • fix features and data args
    patil-suraj committed 4 years ago
  • main
    patil-suraj committed 4 years ago
  • add requirements
    patil-suraj committed 4 years ago
  • add column name args
    patil-suraj committed 4 years ago
  • fix captions
    patil-suraj committed 4 years ago
  • don't jit transforms
    patil-suraj committed 4 years ago
  • fix caption
    patil-suraj committed 4 years ago
  • fix labels, handle attention mask
    patil-suraj committed 4 years ago
  • convert pixel values to numpy
    patil-suraj committed 4 years ago
  • labels => input_ids
    patil-suraj committed 4 years ago
  • transform images on the fly
    patil-suraj committed 4 years ago
  • use AutoModel class, create the hybird model outside of the script
    patil-suraj committed 4 years ago
  • fix version message
    patil-suraj committed 4 years ago
  • add readme
    patil-suraj committed 4 years ago
  • Apply suggestions from code review
    patil-suraj committed 4 years ago
  • adderss review comments
    patil-suraj committed 4 years ago
  • Merge branch 'clip-train-script' of https://github.com/patil-suraj/transformers into clip-train-script
    patil-suraj committed 4 years ago
  • add more comments
    patil-suraj committed 4 years ago
  • allow freezing vision and text models
    patil-suraj committed 4 years ago
FAQ Terms Privacy Refunds Impressum
Loading