We provide several helpful scripts to prepare data, to convert VISSL models to detectron2 compatible models or to convert caffe2 models to VISSL compatible models.

The following scripts are optional as VISSL's dataset_catalog.py supports reading the downloaded data directly for most of these. For all the datasets below, we assume that the datasets are in the format as described here.

python extra_scripts/create_coco_data_files.py \
    --json_annotations_dir /path/to/coco/annotations/ \
    --output_dir /tmp/vissl/datasets/coco/ \
    --train_imgs_path /path/to/coco/train2014 \
    --val_imgs_path /path/to/coco/val2014

0. for VOC2007 data_source_dir='/mnt/fair/VOC2007/' output_dir='/path/to/my/output/dir/voc2007/'

1. for VOC2012 data_source_dir='/mnt/fair/VOC2012/' output_dir='/path/to/my/output/dir/voc2012'

• For VOC2007 dataset:

python extra_scripts/create_voc_data_files.py \
    --data_source_dir /path/to/VOC2007/ \
    --output_dir /tmp/vissl/datasets/voc07/

• For VOC2012 dataset:

python extra_scripts/create_voc_data_files.py \
    --data_source_dir /path/to/VOC2012/ \
    --output_dir /tmp/vissl/datasets/voc12/

python extra_scripts/create_imagenet_data_files.py \
    --data_source_dir /path/to/imagenet_full_size/ \
    --output_dir /tmp/vissl/datasets/imagenet1k/

python extra_scripts/create_imagenet_data_files.py \
    --data_source_dir /path/to/places205/ \
    --output_dir /tmp/vissl/datasets/places205/

python extra_scripts/create_imagenet_data_files.py \
    --data_source_dir /path/to/places365/ \
    --output_dir /tmp/vissl/datasets/places365/

Low-shot image classification is one of the benchmark tasks in the paper. VISSL support low-shot sampling and benchmarking on the PASCAL VOC dataset only.

We train on trainval split of VOC2007 dataset which has 5011 images and 20 classes. Hence the labels are of shape 5011 x 20. We generate 5 independent samples (for a given low-shot value k) by essentially generating 5 independent target files. For each class, we randomly pick the positive k samples and 19 * k negatives. Rest of the samples are ignored. We perform low-shot image classification on various different layers on the model (AlexNet, ResNet-50). The targets targets_data_file is usually obtained by extracting features for a given layer. Below command generates 5 samples for various k values:

python extra_scripts/create_voc_low_shot_samples.py \
    --targets_data_file /path/to/voc/numpy_targets.npy \
    --output_path /tmp/vissl/datasets/voc07/low_shot/labels/ \
    --k_values "1,2,4,8,16,32,64,96" \
    --num_samples 5

We provide scripts to change problem complexity of Jigsaw approach (as an axis of scaling in paper).

For the problem of Jigsaw, we vary the number of permutations used to solve the jigsaw task. In the paper, the permutations used ∈ [100, 2000, 10000]. We provide these permutations files for download here. To generate the permutations, use the command below:

python extra_scripts/generate_jigsaw_permutations.py \
    --output_dir /tmp/vissl//jigsaw_perms/ \
    -- N 2000

We provide scripts to convert VISSL models to Detectron2 and ClassyVision compatible models.

All the ResNe(X)t models in VISSL can be converted to Detectron2 weights using following command:

python extra_scripts/convert_vissl_to_detectron2.py \
    --input_model_file <input_model>.pth  \
    --output_model <d2_model>.torch \
    --weights_type torch \
    --state_dict_key_name classy_state_dict

All the ResNe(X)t models in VISSL can be converted to Detectron2 weights using following command:

python extra_scripts/convert_vissl_to_classy_vision.py \
    --input_model_file <input_model>.pth  \
    --output_model <d2_model>.torch \
    --state_dict_key_name classy_state_dict

All the ResNe(X)t models in VISSL can be converted to Torchvision weights using following command:

python extra_scripts/convert_vissl_to_torchvision.py \
    --model_url_or_file <input_model>.pth  \
    --output_dir /path/to/output/dir/ \
    --output_name <my_converted_model>.torch

We provide conversion of all the caffe2 models in the paper

All the models have been added to ICCV19_MODEL_ZOO_FB.md.

Jigsaw model:

python extra_scripts/convert_caffe2_to_torchvision_resnet.py \
    --c2_model <model>.pkl \
    --output_model <pth_model>.torch \
    --jigsaw True --bgr2rgb True

Colorization model:

python extra_scripts/convert_caffe2_to_torchvision_resnet.py \
    --c2_model <model>.pkl \
    --output_model <pth_model>.torch \
    --bgr2rgb False

Supervised model:

python extra_scripts/convert_caffe2_to_pytorch_rn50.py \
    --c2_model <model>.pkl \
    --output_model <pth_model>.torch \
    --bgr2rgb True

AlexNet Jigsaw models:

python extra_scripts/convert_caffe2_to_vissl_alexnet.py \
    --weights_type caffe2 \
    --model_name jigsaw \
    --bgr2rgb True \
    --input_model_weights <model.pkl> \
    --output_model <pth_model>.torch

AlexNet Colorization models:

python extra_scripts/convert_caffe2_to_vissl_alexnet.py \
    --weights_type caffe2 \
    --model_name colorization \
    --input_model_weights <model.pkl> \
    --output_model <pth_model>.torch

AlexNet Supervised models:

python extra_scripts/convert_caffe2_to_vissl_alexnet.py \
    --weights_type caffe2 \
    --model_name supervised \
    --bgr2rgb True \
    --input_model_weights <model.pkl> \
    --output_model <pth_model>.torch

We provide scripts to convert ClassyVision models to VISSL compatible models.

python extra_scripts/convert_classy_vision_to_vissl_resnet.py \
    --input_model_file <input_model>.pth  \
    --output_model <d2_model>.torch \
    --depth 50

AlexNet RotNet model:

python extra_scripts/convert_caffe2_to_vissl_alexnet.py \
    --weights_type torch \
    --model_name rotnet \
    --input_model_weights <model> \
    --output_model <pth_model>.torch

AlexNet DeepCluster model:

python extra_scripts/convert_alexnet_models.py \
    --weights_type torch \
    --model_name deepcluster \
    --input_model_weights <model> \
    --output_model <pth_model>.torch