v2.5.1

• 🗃️ Add full datasets e7c9ae2
• 🚀 Allow usage of Windows (in particular dealing with absence of JAX) d29aac9
• 🔧 pixel_classes_cond is empty list by default 33819a2

v2.5.0

• ✨ Add post-processing (notebook and functions) a780b0e 14ac94f
• ✨ Add pixel_classes config parameter b33f580 4f12f47
• 🚀 metrics.evaluate returns now generated data and corresponding indicators 5043a8b

v2.4.0

• 📝 Add configuration summary 6fed0e4
• 🚀 Increase the number of normalization iteration to 2 281301d

v2.3.2

• 🐛 Pad one_hot_data if sample has not all the classes in sample_pixels_2d_np 6ad33ac
• 📝 Add example images from Rongier, 2016 and Stanford deltas f25a407

v2.3.1

• 🚀 Cast to bool in one-hot function to save memory 64a4ef4
• ✨ Add asymetric attention 3028bd6
• ✨ Add conditional visualizer 767f9a5

v2.2.0

• 🏗️ Move files and folders of the lib in gan_facies folder 11b06ea
• 🏗️ Update architecture of tests 2cccc6c
• 🚚 Move github_actions in root 9b5417a

v2.1.1

• 🚀 Allow symmetric attention structure in generator and discriminator e122be0
• ✨ Add number of parameters at the beginning of the training 0c278b7
• ✨ Add MACs computation during evaluation 9ece448

v2.1.0

• ✨ Add probability map for conditional models 621e85c 5cd554c
• 🤡 Mock some libs to make more unitary tests d3a5c6b

v2.0.2

• ✨ Add more customization in SelfAttention 00e7123
• 🤡 Mock some 3rd party code in tests 00901f2 2dce95e
• ✨ Add separate function in utils/metrics/tools.py to get reference indicators for metric computation 790c320

v2.0.1

• ✨ Add conditional accuracy 77f4acc
• ✨ Add feature to increase the size of the sampled pixel (uniform class) for conditional models 191b877
• 📈 Add white grid in sampled image grids to separate the images 45df6f8