Visualization¶

Visualization

Learning Rate Schedule Visualization¶

python tools/visualizations/vis_lr.py \
    ${CONFIG_FILE} \
    --dataset-size ${DATASET_SIZE} \
    --ngpus ${NUM_GPUs}
    --save-path ${SAVE_PATH} \
    --title ${TITLE} \
    --style ${STYLE} \
    --window-size ${WINDOW_SIZE}
    --cfg-options

Description of all arguments：

config : The path of a model config file.
dataset-size : The size of the datasets. If set，build_dataset will be skipped and ${DATASET_SIZE} will be used as the size. Default to use the function build_dataset.
ngpus : The number of GPUs used in training, default to be 1.
save-path : The learning rate curve plot save path, default not to save.
title : Title of figure. If not set, default to be config file name.
style : Style of plt. If not set, default to be whitegrid.
window-size: The shape of the display window. If not specified, it will be set to 12*7. If used, it must be in the format 'W*H'.
cfg-options : Modifications to the configuration file, refer to Tutorial 1: Learn about Configs.

Note

Loading annotations maybe consume much time, you can directly specify the size of the dataset with dataset-size to save time.

Examples：

python tools/visualizations/vis_lr.py configs/classification/imagenet/resnet/resnet50_4xb64_step_ep100.py

When using ImageNet, directly specify the size of ImageNet, as below:

python tools/visualizations/vis_lr.py configs/classification/imagenet/resnet/resnet50_4xb64_step_ep100.py --dataset-size 1281167 --ngpus 4 --save-path ./resnet50_4xb64_step_ep100.jpg

(back to top)

Class Activation Map Visualization¶

OpenMixup provides tools\visualizations\vis_cam.py tool to visualize class activation map. Please use pip install "grad-cam>=1.3.6" command to install pytorch-grad-cam. The implementation is modified according to MMClassification (thanks to their contributions).

The supported methods are as follows:

Method	What it does
GradCAM	Weight the 2D activations by the average gradient
GradCAM++	Like GradCAM but uses second order gradients
XGradCAM	Like GradCAM but scale the gradients by the normalized activations
EigenCAM	Takes the first principle component of the 2D Activations (no class discrimination, but seems to give great results)
EigenGradCAM	Like EigenCAM but with class discrimination: First principle component of Activations*Grad. Looks like GradCAM, but cleaner
LayerCAM	Spatially weight the activations by positive gradients. Works better especially in lower layers

Command：

python tools/visualizations/vis_cam.py \
    ${IMG} \
    ${CONFIG_FILE} \
    ${CHECKPOINT} \
    [--target-layers ${TARGET-LAYERS}] \
    [--preview-model] \
    [--method ${METHOD}] \
    [--target-category ${TARGET-CATEGORY}] \
    [--save-path ${SAVE_PATH}] \
    [--vit-like] \
    [--num-extra-tokens ${NUM-EXTRA-TOKENS}]
    [--aug_smooth] \
    [--eigen_smooth] \
    [--device ${DEVICE}] \
    [--cfg-options ${CFG-OPTIONS}]

Description of all arguments：

img : The target picture path.
config : The path of the model config file.
checkpoint : The path of the checkpoint.
--target-layers : The target layers to get activation maps, one or more network layers can be specified. If not set, use the norm layer of the last block.
--preview-model : Whether to print all network layer names in the model.
--method : Visualization method, supports GradCAM, GradCAM++, XGradCAM, EigenCAM, EigenGradCAM, LayerCAM, which is case insensitive. Defaults to GradCAM.
--target-category : Target category, if not set, use the category detected by the given model.
--save-path : The path to save the CAM visualization image. If not set, the CAM image will not be saved.
--vit-like : Whether the network is ViT-like network.
--num-extra-tokens : The number of extra tokens in ViT-like backbones. If not set, use num_extra_tokens the backbone.
--aug_smooth : Whether to use TTA(Test Time Augment) to get CAM.
--eigen_smooth : Whether to use the principal component to reduce noise.
--device : The computing device used. Default to ‘cpu’.
--cfg-options : Modifications to the configuration file, refer to Tutorial 1: Learn about Configs.

Note

The argument --preview-model can view all network layers names in the given model. It will be helpful if you know nothing about the model layers when setting --target-layers.

Examples(CNN)：

Here are some examples of target-layers in ResNet-50, which can be any module or layer:

'backbone.layer4' means the output of the forth ResLayer.
'backbone.layer4.2' means the output of the third BottleNeck block in the forth ResLayer.
'backbone.layer4.2.conv1' means the output of the conv1 layer in above BottleNeck block.

Note

For ModuleList or Sequential, you can also use the index to specify which sub-module is the target layer.

For example, the backbone.layer4[-1] is the same as backbone.layer4.2 since layer4 is a Sequential with three sub-modules.

Use different methods to visualize CAM for ResNet50, the target-category is the predicted result by the given checkpoint, using the default target-layers.

python tools/visualizations/vis_cam.py \
    demo/bird.JPEG \
    configs/classification/imagenet/resnet/resnet50_4xb64_step_ep100.py \
    https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_batch256_imagenet_20200708-cfb998bf.pth \
    --method GradCAM
    # GradCAM++, XGradCAM, EigenCAM, EigenGradCAM, LayerCAM

Image	GradCAM	GradCAM++	EigenGradCAM	LayerCAM

Use different target-category to get CAM from the same picture. In ImageNet dataset, the category 238 is ‘Greater Swiss Mountain dog’, the category 281 is ‘tabby, tabby cat’.

python tools/visualizations/vis_cam.py \
    demo/cat-dog.png \
    configs/classification/imagenet/resnet/resnet50_4xb64_step_ep100.py \
    https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_batch256_imagenet_20200708-cfb998bf.pth \
    --target-layers 'backbone.layer4.2' \
    --method GradCAM \
    --target-category 238
    # --target-category 281

Category	Image	GradCAM	XGradCAM	LayerCAM
Dog
Cat

Use --eigen-smooth and --aug-smooth to improve visual effects.

python tools/visualizations/vis_cam.py \
    demo/bird.JPEG \
    configs/classification/imagenet/resnet/resnet50_4xb64_step_ep100.py \
    https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_batch256_imagenet_20200708-cfb998bf.pth \
    --target-layers 'backbone.layer4.2' \
    --method LayerCAM \
    --eigen-smooth --aug-smooth

Image	LayerCAM	eigen-smooth	aug-smooth	eigen&aug

Examples(Transformer)：

Here are some examples:

'backbone.norm3' for Swin-Transformer;
'backbone.layers[-1].ln1' for ViT;

For ViT-like networks, such as ViT, T2T-ViT and Swin-Transformer, the features are flattened. And for drawing the CAM, we need to specify the --vit-like argument to reshape the features into square feature maps.

Besides the flattened features, some ViT-like networks also add extra tokens like the class token in ViT and T2T-ViT, and the distillation token in DeiT. In these networks, the final classification is done on the tokens computed in the last attention block, and therefore, the classification score will not be affected by other features and the gradient of the classification score with respect to them, will be zero. Therefore, you shouldn’t use the output of the last attention block as the target layer in these networks.

To exclude these extra tokens, we need know the number of extra tokens. Almost all transformer-based backbones in MMClassification have the num_extra_tokens attribute. If you want to use this tool in a new or third-party network that don’t have the num_extra_tokens attribute, please specify it the --num-extra-tokens argument.

Visualize CAM for Swin Transformer, using default target-layers:

python tools/visualizations/vis_cam.py \
    demo/bird.JPEG \
    configs/classification/imagenet/swin_transformer/swin_tiny_8xb128_fp16_ep300.py \
    https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_tiny_224_b16x64_300e_imagenet_20210616_090925-66df6be6.pth \
    --vit-like

Visualize CAM for Vision Transformer(ViT):

python tools/visualizations/vis_cam.py \
    demo/bird.JPEG \
    configs/classification/imagenet/deit/deit_base_8xb128_ep300.py \
    https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_pt-16xb64_in1k_20211124-6f40c188.pth \
    --vit-like \
    --target-layers 'backbone.layers[-1].ln1'

Visualize CAM for T2T-ViT:

python tools/visualizations/vis_cam.py \
    demo/bird.JPEG \
    configs/classification/imagenet/t2t_vit/t2t_vit_tiny_14_ema_8xb64_ep310.py \
    https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_3rdparty_8xb64_in1k_20210928-b7c09b62.pth \
    --vit-like \
    --target-layers 'backbone.encoder[-1].ln1'

Image	ResNet50	ViT	Swin	T2T-ViT

(back to top)

Loss Landscape Visualization¶

OpenMixup provides tools\visualizations\vis_loss_landscape.py tool to visualize loss landscapes of classification models. Please use pip install h5py command to install h5py. The implementation is borrowed and modified from loss-landscape in Visualizing the Loss Landscape of Neural Nets. Thanks to their contributions.

We run vis_loss_landscape.py by dist_vis_loss.sh as DDP testing. This tool calculates and visualizes the loss surface along random direction(s) near the optimal parameters in parallel with multiple GPUs. To avoid the blocking error from h5py, you can use export HDF5_USE_FILE_LOCKING=FALSE before running the visualization experiment.

Command

bash tools/visualizations/dist_vis_loss.sh \
    ${CONFIG_FILE} \
    ${GPUS} \
    ${CHECKPOINT} \
    [--plot_mode ${surface | trajectory | surface+trajectory}] \
    [--dir_type ${weights | states}] \
    [--dir_type ${weights | states}] \
    [--x ${xmin:x_max:xnum}] \
    [--y ${ymin:y_max:ynum}] \
    [--xnorm ${filter | layer | weight}] \
    [--ynorm ${filter | layer | weight}] \
    [--xignore ${biasbn}] \
    [--yignore ${biasbn}] \
    [--cfg-options ${CFG-OPTIONS}]

Description of all arguments：

config : The path of the model config file.
gpu : The number of GPUs to evaluate the model.
checkpoint : The path of the checkpoint.
--plot_mode : The plot mode of loss landscape (‘surface’ | ‘trajectory’ | ‘surface+trajectory’).
--model_file2 : Using (model_file2 - model_file1) as the xdirection.
--model_file3 : Using (model_file3 - model_file1) as the ydirection.
--dir_file : Specify the name of direction file, or the path to a direction file.
--dir_type : The plotting direction type (‘weights’ | ‘states’). Note that ‘states’ indicates the direction contains dimensions for all parameters as well as the statistics of the BN layers (running_mean and running_var), while ‘weights’ indicates the direction has the same dimensions as the learned parameters, including bias and parameters in the BN layers.
--x : A string with format (‘xmin:x_max:xnum’).
--y : A string with format (‘ymin:ymax:ynum’).
--xnorm : The direction normalization (‘filter’ | ‘layer’ | ‘weight’).
--ynorm : The direction normalization (‘filter’ | ‘layer’ | ‘weight’).
--xignore : To ignore the direction corresponding to bias and BN parameters (fill the corresponding entries in the random vector with zeros) in the x-axis (“biasbn”).
--yignore : To ignore bias and BN parameters the y-axis (“biasbn”).
--surf_file : The customize the name of surface file, could be an existing file.
--proj_file : The .h5 file contains projected optimization trajectory.
--loss_max : Maximum value to show in 1D plot.
--vmax : Maximum value to map in the plot.
--vmin : Miminum value to map in the plot.
--vlevel : Plot contours every vlevel.
--log : Whether to use log scale for loss values.
--plot_format : The save format of plotted matplotlib images (defaults to “png”).
--model_folder : Folders for models to be projected (defaults to work_dirs).
--prefix : The prefix for the checkpint model for plotting the trajectory (defaults to “epoch”).
--start_epoch : The min index of epochs for plotting the trajectory.
--max_epoch : The max number of epochs for plotting the trajectory.
--save_interval : The interval to save models for plotting the trajectory.
--cfg-options : Modifications to the configuration file, refer to Tutorial 1: Learn about Configs.

Examples：

Visualizing 1D linear interpolations. The 1D linear interpolation method evaluates the loss values along the direction between two minimizers of the same network loss function.
```
bash tools/visualizations/dist_vis_loss.sh \
    configs/classification/cifar100/wa/resnet18_CE_bs100.py 1 ${CHECKPOINT} \
    --plot_mode surface --x=-1:1:51 --dir_type weights --xnorm filter --xignore biasbn
```
Note

--x=-1:1:51 sets the range and resolution for the plot. The x-coordinates in the plot will run from -1 to 1 (the minimizers are located at 0 and 1), and the loss value will be evaluated at 51 locations along this line.

--dir_type weights indicates the direction has the same dimensions as the learned parameters, including bias and parameters in the BN layers.

--xnorm filter normalizes the random direction at the filter level. Here, a “filter” refers to the parameters that produce a single feature map. For fully connected layers, a “filter” contains the weights that contribute to a single neuron.

--xignore biasbn ignores the direction corresponding to bias and BN parameters (fill the corresponding entries in the random vector with zeros).

Visualizing 2D loss contours and 3D loss surfaces. To plot the loss landscape, we choose two random directions and normalize them in the same way as the 1D plotting.

bash tools/visualizations/dist_vis_loss.sh \
    configs/classification/cifar100/wa/resnet18_CE_bs100.py 1 ${CHECKPOINT} \
    --plot_mode surface --x=-1:1:51 --y=-1:1:51 --dir_type weights \
    --xnorm filter --xignore biasbn --ynorm filter --yignore biasbn --vlevel 0.1

2D contours

3D surfaces

Visualizing the loss trajectory in 2D loss contours. We plot the loss surfaces and the optimization trajectory in a 2D plot, which requires a list of models (saved in the path to work_dirs).

bash tools/visualizations/dist_vis_loss.sh \
    configs/classification/cifar100/wa/resnet18_CE_bs100.py 1 ${CHECKPOINT} \
    --plot_mode surface+trajectory --x=-1:1:51 --y=-1:1:51 --dir_type weights \
    --xnorm filter --xignore biasbn --ynorm filter --yignore biasbn \
    --start_epoch 0 --max_epoch 200 --save_interval 10

FAQs¶

None

(back to top)