收集整理了一些常见的可以用到的深度学习网址、工具

1. 预训练模型

https://github.com/Cadene/pretrained-models.pytorch

https://github.com/rwightman/pytorch-image-models

https://github.com/welkin-feng/ComputerVision

2. 数据增强

https://github.com/albumentations-team/albumentations

3. 标记工具

Labelme: Image Polygonal Annotation with Python

LabelImg：LabelImg is a graphical image annotation tool and label object bounding boxes in images

4. 数据集查找

! ! ! You can find datasets in Paper Beachmark

Kaggle

Google Datasets Search Engine

Microsoft Datasets

Computer Vision Datasets

Github awesomedata

UCI Machine Learning Repository.

Amazon Datasets

Government Datasets: EU US NZL IND

5. 模型分析工具

(1) 卷积层输出大小计算

https://ezyang.github.io/convolution-visualizer/index.html

(2) 计算模型参数量

https://github.com/sksq96/pytorch-summary

(3) 模型可视化工具

Netron: now supports ONNX, Keras, CoreML, Caffe2, Mxnet, Pytorch and Tensorflow.

Graphviz: Pytorch

6. 可视化工具

visdom

# Example using Visdom.
vis = visdom.Visdom(env='Learning curve', use_incoming_socket=False)
assert self._visdom.check_connection()
self._visdom.close()
options = collections.namedtuple('Options', ['loss', 'acc', 'lr'])(
    loss={'xlabel': 'Epoch', 'ylabel': 'Loss', 'showlegend': True},
    acc={'xlabel': 'Epoch', 'ylabel': 'Accuracy', 'showlegend': True},
    lr={'xlabel': 'Epoch', 'ylabel': 'Learning rate', 'showlegend': True})

for t in epoch(80):
    tran(...)
    val(...)
    vis.line(X=torch.Tensor([t + 1]), Y=torch.Tensor([train_loss]),
             name='train', win='Loss', update='append', opts=options.loss)
    vis.line(X=torch.Tensor([t + 1]), Y=torch.Tensor([val_loss]),
             name='val', win='Loss', update='append', opts=options.loss)
    vis.line(X=torch.Tensor([t + 1]), Y=torch.Tensor([train_acc]),
             name='train', win='Accuracy', update='append', opts=options.acc)
    vis.line(X=torch.Tensor([t + 1]), Y=torch.Tensor([val_acc]),
             name='val', win='Accuracy', update='append', opts=options.acc)
    vis.line(X=torch.Tensor([t + 1]), Y=torch.Tensor([lr]),
             win='Learning rate', update='append', opts=options.lr)

Tensorboard

• acc / loss

from tensorboard import SummaryWriter
writer = SummaryWriter()
for n_iter in range(100):
    dummy_s1 = torch.rand(1)
    writer.add_scalar('data/scalar1', dummy_s1[0], n_iter)
writer.close()

• img

from tensorboard import SummaryWriter
import torchvision.utils as vutils
writer = SummaryWriter()
if n_iter % 10 == 0:
    x = vutils.make_grid(dummy_img, normalize=True, scale_each=True)
    writer.add_image('Image', x, n_iter)
writer.close()

• 在一张图中加入两条曲线

for i in range(100):
    writer.add_scalars('run_14h', {'xsinx':i*np.sin(i/r),
                                    'xcosx':i*np.cos(i/r),
                                    'tanx': np.tan(i/r)}, i)

7. Pytorch 加速

NVIDIA/DLAI: https://github.com/NVIDIA/DALI

Efficient-Pytorch: https://github.com/Lyken17/Efficient-PyTorch

NVIDIA/APEX: https://github.com/nvidia/apex

8. 性能分析工具

• nvidia-smi
• htop
• iotop
• nvtop
• py-spy
• strace

9. 深度学习绘图

• ML Visuals
• PlotNeuralNet

10. 其他辅助工具

• byobu
• screen