0.1.0

What's Changed

[0.1.0] - 2024-02-23

Fixed

• Fix the error of converting tuple to list.
• Fix the problem of apparent memory or non-release after training failure.
• Repair nan problems in training
• Fix the error of saving D model and the prompt to increase the empty dataset

Added

• Provide pre-training weight, which is comparable to the official effect.
• Simplify the training process and use yaml instead of py files

Full Changelog: https://github.com/Lornatang/Real_ESRGAN-PyTorch/commits/0.1.0