Age | Commit message (Collapse) | Author | |
---|---|---|---|
2020-12-29 | Return canonical features at condition 1 for later aggregation | Jordan Gong | |
2020-12-28 | Wrap the auto-encoder, return 3 losses at t2 | Jordan Gong | |
2020-12-27 | Fix inconsistency and API deprecation issues in decoder | Jordan Gong | |
1. Add default output channels of decoder 2. Replace deprecated torch.nn.functional.sigmoid with torch.sigmoid | |||
2020-12-27 | Refine auto-encoder | Jordan Gong | |
1. Wrap fully connected layers 2. Introduce hyperparameter tuning in constructor | |||
2020-12-24 | Optimize imports | Jordan Gong | |
2020-12-24 | Change the usage of layers and reorganize relations of layers | Jordan Gong | |
1. Add batch normalization and activation to layers 2. VGGConv2d and FocalConv2d inherits to BasicConv2d; DCGANConvTranspose2d inherits to BasicConvTranspose2d | |||
2020-12-23 | Modify activation functions after conv or trans-conv in auto-encoder | Jordan Gong | |
1. Make activation functions be inplace ops 2. Change Leaky ReLU to ReLU in decoder | |||
2020-12-23 | Refactor and refine auto-encoder | Jordan Gong | |
1. Wrap Conv2d 3x3-padding-1 to VGGConv2d 2. Wrap ConvTranspose2d 4x4-stride-4-padding-1 to DCGANConvTranspose2d 3. Turn off bias in conv since the employment of batch normalization | |||
2020-12-23 | Reshape feature before decode | Jordan Gong | |
2020-12-23 | Split modules to different files | Jordan Gong | |