luzai luzai
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| if args.inp_gw != 0: | |
| input_imgs_grad_attached = torch.autograd.grad( | |
| outputs=loss, inputs=imgs, | |
| create_graph=True, retain_graph=True, | |
| only_inputs=True | |
| )[0] | |
| input_imgs_grad_attached = input_imgs_grad_attached.view( | |
| input_imgs_grad_attached.size(0), -1 | |
| ) | |
| if args.aux == 'l1_grad': |
luzai
/ test_superkernel.py
Last active
June 12, 2019 03:23
this version implementation of superkernel tend to gradient explode.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import torch | |
| import torch.nn as nn | |
| import torch.nn.functional as F | |
| from easydict import EasyDict | |
| import numpy as np | |
| __all__ = ['singlepath'] | |
| conf = EasyDict() | |
| conf.conv2dmask_drop_ratio = .0 |
luzai
/ get_layer_bug.py
Last active
May 8, 2017 03:06
Try to fix shapes bug in keras/example/mnist_net2net.py
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| '''This is an implementation of Net2Net experiment with MNIST in | |
| 'Net2Net: Accelerating Learning via Knowledge Transfer' | |
| by Tianqi Chen, Ian Goodfellow, and Jonathon Shlens | |
| arXiv:1511.05641v4 [cs.LG] 23 Apr 2016 | |
| http://arxiv.org/abs/1511.05641 | |
| Notes | |
| - What: | |
| + Net2Net is a group of methods to transfer knowledge from a teacher neural |