tools

Tools

The tools/ folder contains some useful tools for developing and experimenting. It is not guaranteed that these tools will work for all kind of use-cases, datasets and datatypes but even then they can be used as a starting point and can be adapted with a few changes. These scripts support the general Hydra override syntax as described in config, together with some additional arguments (these can be used via the argparse syntax with a -- prefix). For each tool use the -h flag (e.g. python tools/show_data.py -h) to see all additional options.

show_data.py

Load and Visualize the pytorch dataset which is defined in the dataset config. Can be used for data inspection and to view different data augmentation pipelines.

• dataset: Name of the dataset config (see here)
• --augmentation: Which augmentations to use: None (by default), train, val or test. If A.Normalization(std=...,mean=...) is part of the augmentations, this will be undone during visualization to get a better interpretable image
• --split: which split to use: train, val or test Dataset (train by default)
• --segmentation: type of segmentation - semantic or instance, depending on the dataset

pyhton tools/show_data.py dataset=<dataset.name>
pyhton tools/show_data.py dataset=Cityscapes --split=val --augmentation=val

show_prediction.py

Show the predictions of a trained model. Basically has the same syntax as the validation/testing, but visualizes the result instead of calculating metrics. The Test Dataset is used for predicting by default together with the train data-augmentations. For a nicer appearance the normalization operation is undone during visualization (not for prediction). (Note, depending on the size of the input, the inference time of the model and the available hardware, there might be a delay when sliding through the images)

• ckpt_dir: path to the checkpoint which should be used
• --augmentation: Which augmentations to use: train, val or test (by default). If A.Normalization(std=...,mean=...) is part of the augmentations, this will be undone during visualization to get a better interpretable image
• --split: which split to use: train, val or test Dataset (train by default)
• --segmentation: type of segmentation - semantic or instance, depending on the dataset

python tools/show_prediction.py ckpt_dir=<path>
python tools/show_prediction.py ckpt_dir=ckpt_dir="/../Semantic_Segmentation/logs/VOC2010_Context/hrnet/baseline_/2022-02-15_13-51-42" --split=test --augmentation=test

dataset_stats.py

Getting some basic stats and visualizations about the dataset like: mean and std for each channel, appearances and ratio of classes and potential class weights. Also, the color encoding of classes is visualized (can be useful for the show_data or show_prediction scripts) The output will be saved in dataset_stats/dataset_name/. If no output is wanted use the supress_output flag

• dataset: Name of the dataset config (see here)
• --name: Prefix for naming the results
• --split: which split to use: train, val or test
• --img_only: Flag for only analyse Image Data
• --mask_only: Flag for only analyse Mask Data
• --supress_output: Flag to supress creation of an output directory and output files

  python tools/dataset_stats.py dataset=<dataset.name>
  pyhton tools/dataset_stats.py dataset=Cityscapes

lr_finder.py

Implementation to use pytorch lightning's Learning Rate Finder to get some guidance when choosing an optimal initial lr (Should be used with caution, especially if random augmentations are used).

• --num_training_samples: number of batches which are used from the lr finder (100 by default)

python tools/lr_finder.py
pyhton tools/lr_finder.py dataset=Cityscapes model=hrnet --num_training=300

Acknowledgements

    

This Repository is developed and maintained by the Applied Computer Vision Lab (ACVL) of Helmholtz Imaging.