1.0.6.2
- Update TF backend
- Add support to tensorflow 2.0 (both legacy and eager mode)
- Refactor torch backend models
- Add
--caching_datasetto cache transformed data into memory (ignored whenmemory_limitset). - Fix FastMetrics multi-threads issue
- Fix loading issue when inferring with VSR models
- Fix bugs of DataLoader #108, #109
- Fix an error that dataloader may mess up the file order
- Add a dataset checker to help verify the DDF
- Fix a bug of memory_limit issue
- Add --export to Train/eval.py
- Fix wrong data format in ImageProcess
- Google Drive download api expired, fallback to a link message
- Add SRMD (CVPR 2018) by Kai Zhang et. al.
- Upload to PyPI.
- Trace changelog since 1.0.0
- Move VSRTorch package into VSR
- Use
~/.vsr/config.yml(C:\Users\<user>\.vsr\config.ymlfor windows) to config the package - Provide common API for users to add into their own project easily
- Use
- Refresh detailed documents