torch.nn.ReLU is the recommended activation module. removed the custom defined module...
authorManu Mathew <a0393608@ti.com>
Thu, 12 Mar 2020 10:45:36 +0000 (16:15 +0530)
committerManu Mathew <a0393608@ti.com>
Thu, 12 Mar 2020 10:52:15 +0000 (16:22 +0530)
commitab84ec9f9e27a32f45a2e6372f32ba3fe9b1a799
treea449c358b2c6732a83d07551f483f39b685a6a6b
parentdaeea4b13859b5a45038d57eabd3992bb6297125
torch.nn.ReLU is the recommended activation module. removed the custom defined module called ReLUN - if fixed range activation module is needed torch.nn.Hardtanh can be used.
16 files changed:
modules/pytorch_jacinto_ai/engine/train_pixel2pixel.py
modules/pytorch_jacinto_ai/vision/losses/loss_utils.py
modules/pytorch_jacinto_ai/vision/losses/norm_loss.py
modules/pytorch_jacinto_ai/vision/models/multi_input_net.py
modules/pytorch_jacinto_ai/vision/models/pixel2pixel/deeplabv3lite.py
modules/pytorch_jacinto_ai/vision/transforms/__init__.py
modules/pytorch_jacinto_ai/vision/transforms/image_transform_utils.py
modules/pytorch_jacinto_ai/vision/transforms/image_transforms.py
modules/pytorch_jacinto_ai/vision/transforms/image_transforms_xv12.py [new file with mode: 0644]
modules/pytorch_jacinto_ai/xnn/layers/activation.py
modules/pytorch_jacinto_ai/xnn/layers/common_blocks.py
modules/pytorch_jacinto_ai/xnn/layers/layer_config.py
modules/pytorch_jacinto_ai/xnn/quantize/quant_graph_module.py
modules/pytorch_jacinto_ai/xnn/quantize/quant_train_module.py
modules/pytorch_jacinto_ai/xnn/utils/load_weights.py
modules/pytorch_jacinto_ai/xnn/utils/module_utils.py