Modelzoo¶

Action Localization Models¶

BMN¶

Introduction¶

@inproceedings{lin2019bmn,
  title={Bmn: Boundary-matching network for temporal action proposal generation},
  author={Lin, Tianwei and Liu, Xiao and Li, Xin and Ding, Errui and Wen, Shilei},
  booktitle={Proceedings of the IEEE International Conference on Computer Vision},
  pages={3889--3898},
  year={2019}
}

Model Zoo¶

ActivityNet feature¶

config	feature	gpus	pretrain	AR@100	AUC	gpu_mem(M)	iter time(s)	ckpt	log	json
bmn_400x100_9e_2x8_activitynet_feature	cuhk_mean_100	2	None	75.28	67.22	5420	3.27	ckpt	log	json
	mmaction_video	2	None	75.43	67.22	5420	3.27	ckpt	log	json
	mmaction_clip	2	None	75.35	67.38	5420	3.27	ckpt	log	json

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
For feature column, cuhk_mean_100 denotes the widely used cuhk activitynet feature extracted by anet2016-cuhk, mmaction_video and mmaction_clip denote feature extracted by mmaction, with video-level activitynet finetuned model or clip-level activitynet finetuned model respectively.

For more details on data preparation, you can refer to ActivityNet feature in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train BMN model on ActivityNet features dataset.

python tools/train.py configs/localization/bmn/bmn_400x100_2x8_9e_activitynet_feature.py

For more details and optional arguments infos, you can refer to Training setting part in getting_started .

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test BMN on ActivityNet feature dataset.

### Note: If evaluated, then please make sure the annotation file for test data contains groundtruth.
python tools/test.py configs/localization/bmn/bmn_400x100_2x8_9e_activitynet_feature.py checkpoints/SOME_CHECKPOINT.pth --eval AR@AN --out results.json

For more details and optional arguments infos, you can refer to Test a dataset part in getting_started .

BSN¶

Introduction¶

@inproceedings{lin2018bsn,
  title={Bsn: Boundary sensitive network for temporal action proposal generation},
  author={Lin, Tianwei and Zhao, Xu and Su, Haisheng and Wang, Chongjing and Yang, Ming},
  booktitle={Proceedings of the European Conference on Computer Vision (ECCV)},
  pages={3--19},
  year={2018}
}

Model Zoo¶

ActivityNet feature¶

config	feature	gpus	pretrain	AR@100	AUC	gpu_mem(M)	iter time(s)	ckpt	log	json
bsn_400x100_1x16_20e_activitynet_feature	cuhk_mean_100	1	None	74.65	66.45	41(TEM)+25(PEM)	0.074(TEM)+0.036(PEM)	ckpt_tem ckpt_pem	log_tem log_pem	json_tem json_pem
	mmaction_video	1	None	74.93	66.74	41(TEM)+25(PEM)	0.074(TEM)+0.036(PEM)	ckpt_tem ckpt_pem	log_tem log_pem	json_tem json_pem
	mmaction_clip	1	None	75.19	66.81	41(TEM)+25(PEM)	0.074(TEM)+0.036(PEM)	ckpt_tem ckpt_pem	log_tem log_pem	json_tem json_pem

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
For feature column, cuhk_mean_100 denotes the widely used cuhk activitynet feature extracted by anet2016-cuhk, mmaction_video and mmaction_clip denote feature extracted by mmaction, with video-level activitynet finetuned model or clip-level activitynet finetuned model respectively.

For more details on data preparation, you can refer to ActivityNet feature in Data Preparation.

Train¶

You can use the following commands to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Examples:

train BSN(TEM) on ActivityNet features dataset.

python tools/train.py configs/localization/bsn/bsn_tem_400x100_1x16_20e_activitynet_feature.py

train BSN(PEM) on PGM results.

python tools/train.py configs/localization/bsn/bsn_pem_400x100_1x16_20e_activitynet_feature.py

For more details and optional arguments infos, you can refer to Training setting part in getting_started.

Inference¶

You can use the following commands to inference a model.

For TEM Inference

### Note: This could not be evaluated.
python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

For PGM Inference

python tools/bsn_proposal_generation.py ${CONFIG_FILE} [--mode ${MODE}]

For PEM Inference

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Examples:

Inference BSN(TEM) with pretrained model.

python tools/test.py configs/localization/bsn/bsn_tem_400x100_1x16_20e_activitynet_feature.py checkpoints/SOME_CHECKPOINT.pth

Inference BSN(PGM) with pretrained model.

python tools/bsn_proposal_generation.py configs/localization/bsn/bsn_pgm_400x100_activitynet_feature.py --mode train

Inference BSN(PEM) with evaluation metric ‘AR@AN’ and output the results.

### Note: If evaluated, then please make sure the annotation file for test data contains groundtruth.
python tools/test.py configs/localization/bsn/bsn_pem_400x100_1x16_20e_activitynet_feature.py  checkpoints/SOME_CHECKPOINT.pth  --eval AR@AN --out results.json

Test¶

You can use the following commands to test a model.

TEM

### Note: This could not be evaluated.
python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

PGM

python tools/bsn_proposal_generation.py ${CONFIG_FILE} [--mode ${MODE}]

PEM

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Examples:

Test a TEM model on ActivityNet dataset.

python tools/test.py configs/localization/bsn/bsn_tem_400x100_1x16_20e_activitynet_feature.py checkpoints/SOME_CHECKPOINT.pth

Test a PGM model on ActivityNet dataset.

python tools/bsn_proposal_generation.py configs/localization/bsn/bsn_pgm_400x100_activitynet_feature.py --mode test

Test a PEM model with with evaluation metric ‘AR@AN’ and output the results.

python tools/test.py configs/localization/bsn/bsn_pem_400x100_1x16_20e_activitynet_feature.py checkpoints/SOME_CHECKPOINT.pth --eval AR@AN --out results.json

For more details and optional arguments infos, you can refer to Test a dataset part in getting_started.

Action Recognition Models¶

CSN¶

Introduction¶

@inproceedings{inproceedings,
author = {Wang, Heng and Feiszli, Matt and Torresani, Lorenzo},
year = {2019},
month = {10},
pages = {5551-5560},
title = {Video Classification With Channel-Separated Convolutional Networks},
doi = {10.1109/ICCV.2019.00565}
}

@inproceedings{ghadiyaram2019large,
  title={Large-scale weakly-supervised pre-training for video action recognition},
  author={Ghadiyaram, Deepti and Tran, Du and Mahajan, Dhruv},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  pages={12046--12055},
  year={2019}
}

Model Zoo¶

Kinetics-400¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	inference_time(video/s)	gpu_mem(M)	ckpt	log	json
ircsn_ig65m_pretrained_r152_32x2x1_58e_kinetics400_rgb.py	short-side 320	8x4	ResNet152	IG65M	80.14	94.93	x	8517	ckpt	log	json
ircsn_ig65m_pretrained_bnfrozen_r152_32x2x1_58e_kinetics400_rgb.py	short-side 320	8x4	ResNet152	IG65M	82.76	95.68	x	8516	ckpt	log	json

Notes:

The gpus indicates the number of gpu (32G V100) we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8x4 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.
The values in columns named after “reference” are the results got by training on the original repo, using the same model settings.

For more details on data preparation, you can refer to Kinetics400 in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train CSN model on Kinetics-400 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/csn/ircsn_ig65m_pretrained_r152_32x2x1_58e_kinetics400_rgb.py \
    --work-dir work_dirs/ircsn_ig65m_pretrained_r152_32x2x1_58e_kinetics400_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test CSN model on Kinetics-400 dataset and dump the result to a json file.

python tools/test.py configs/recognition/csn/ircsn_ig65m_pretrained_r152_32x2x1_58e_kinetics400_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json --average-clips prob

For more details, you can refer to Test a dataset part in getting_started.

I3D¶

Introduction¶

@inproceedings{inproceedings,
  author = {Carreira, J. and Zisserman, Andrew},
  year = {2017},
  month = {07},
  pages = {4724-4733},
  title = {Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset},
  doi = {10.1109/CVPR.2017.502}
}

@article{NonLocal2018,
  author =   {Xiaolong Wang and Ross Girshick and Abhinav Gupta and Kaiming He},
  title =    {Non-local Neural Networks},
  journal =  {CVPR},
  year =     {2018}
}

Model Zoo¶

Kinetics-400¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	inference_time(video/s)	gpu_mem(M)	ckpt	log	json
i3d_r50_32x2x1_100e_kinetics400_rgb	340x256	8	ResNet50	ImageNet	72.68	90.78	1.7 (320x3 frames)	5170	ckpt	log	json
i3d_r50_32x2x1_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	73.27	90.92	x	5170	ckpt	log	json
i3d_r50_video_32x2x1_100e_kinetics400_rgb	short-side 256p	8	ResNet50	ImageNet	72.85	90.75	x	5170	ckpt	log	json
i3d_r50_dense_32x2x1_100e_kinetics400_rgb	340x256	8x2	ResNet50	ImageNet	72.77	90.57	1.7 (320x3 frames)	5170	ckpt	log	json
i3d_r50_dense_32x2x1_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	73.48	91.00	x	5170	ckpt	log	json
i3d_r50_lazy_32x2x1_100e_kinetics400_rgb	340x256	8	ResNet50	ImageNet	72.32	90.72	1.8 (320x3 frames)	5170	ckpt	log	json
i3d_r50_lazy_32x2x1_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	73.24	90.99	x	5170	ckpt	log	json
i3d_nl_embedded_gaussian_r50_32x2x1_100e_kinetics400_rgb	short-side 256p	8x4	ResNet50	ImageNet	74.71	91.81	x	6438	ckpt	log	json
i3d_nl_gaussian_r50_32x2x1_100e_kinetics400_rgb	short-side 256p	8x4	ResNet50	ImageNet	73.37	91.26	x	4944	ckpt	log	json
i3d_nl_dot_product_r50_32x2x1_100e_kinetics400_rgb	short-side 256p	8x4	ResNet50	ImageNet	73.92	91.59	x	4832	ckpt	log	json

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.

For more details on data preparation, you can refer to Kinetics400 in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train I3D model on Kinetics-400 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/i3d/i3d_r50_32x2x1_100e_kinetics400_rgb.py \
    --work-dir work_dirs/i3d_r50_32x2x1_100e_kinetics400_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test I3D model on Kinetics-400 dataset and dump the result to a json file.

python tools/test.py configs/recognition/i3d/i3d_r50_32x2x1_100e_kinetics400_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json --average-clips prob

For more details, you can refer to Test a dataset part in getting_started.

R2plus1D¶

Introduction¶

@inproceedings{tran2018closer,
  title={A closer look at spatiotemporal convolutions for action recognition},
  author={Tran, Du and Wang, Heng and Torresani, Lorenzo and Ray, Jamie and LeCun, Yann and Paluri, Manohar},
  booktitle={Proceedings of the IEEE conference on Computer Vision and Pattern Recognition},
  pages={6450--6459},
  year={2018}
}

Model Zoo¶

Kinetics-400¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	inference_time(video/s)	gpu_mem(M)	ckpt	log	json
r2plus1d_r34_8x8x1_180e_kinetics400_rgb	short-side 256	8x4	ResNet34	None	67.30	87.65	x	5019	ckpt	log	json
r2plus1d_r34_video_8x8x1_180e_kinetics400_rgb	short-side 256	8	ResNet34	None	67.3	87.8	x	5019	ckpt	log	json
r2plus1d_r34_8x8x1_180e_kinetics400_rgb	short-side 320	8x2	ResNet34	None	68.68	88.36	1.6 (80x3 frames)	5019	ckpt	log	json
r2plus1d_r34_32x2x1_180e_kinetics400_rgb	short-side 320	8x2	ResNet34	None	74.60	91.59	0.5 (320x3 frames)	12975	ckpt	log	json

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.

For more details on data preparation, you can refer to Kinetics400 in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train R(2+1)D model on Kinetics-400 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/r2plus1d/r2plus1d_r34_8x8x1_180e_kinetics400_rgb.py \
    --work-dir work_dirs/r2plus1d_r34_3d_8x8x1_180e_kinetics400_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test R(2+1)D model on Kinetics-400 dataset and dump the result to a json file.

python tools/test.py configs/recognition/r2plus1d/r2plus1d_r34_8x8x1_180e_kinetics400_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json --average-clips=prob

For more details, you can refer to Test a dataset part in getting_started.

SlowFast¶

Introduction¶

@inproceedings{feichtenhofer2019slowfast,
  title={Slowfast networks for video recognition},
  author={Feichtenhofer, Christoph and Fan, Haoqi and Malik, Jitendra and He, Kaiming},
  booktitle={Proceedings of the IEEE international conference on computer vision},
  pages={6202--6211},
  year={2019}
}

Model Zoo¶

Kinetics-400¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	inference_time(video/s)	gpu_mem(M)	ckpt	log	json
slowfast_r50_4x16x1_256e_kinetics400_rgb	short-side 256	8x4	ResNet50	None	74.75	91.73	x	6203	ckpt	log	json
slowfast_r50_video_4x16x1_256e_kinetics400_rgb	short-side 256	8	ResNet50	None	74.34	91.58	x	6203	ckpt	log	json
slowfast_r50_4x16x1_256e_kinetics400_rgb	short-side 320	8x3	ResNet50	None	75.64	92.3	1.6 ((32+4)x10x3 frames)	6203	ckpt	log	json
slowfast_r50_8x8x1_256e_kinetics400_rgb	short-side 256	8x4	ResNet50	None	75.61	92.34	x	9062	ckpt	log	json
slowfast_r50_8x8x1_256e_kinetics400_rgb	short-side 320	8x3	ResNet50	None	76.94	92.8	1.3 ((32+8)x10x3 frames)	9062	ckpt	log	json

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.

For more details on data preparation, you can refer to Kinetics400 in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train SlowFast model on Kinetics-400 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/slowfast/slowfast_r50_4x16x1_256e_kinetics400_rgb.py \
    --work-dir work_dirs/slowfast_r50_4x16x1_256e_kinetics400_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test SlowFast model on Kinetics-400 dataset and dump the result to a json file.

python tools/test.py configs/recognition/slowfast/slowfast_r50_4x16x1_256e_kinetics400_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json --average-clips=prob

For more details, you can refer to Test a dataset part in getting_started.

SlowOnly¶

Introduction¶

@inproceedings{feichtenhofer2019slowfast,
  title={Slowfast networks for video recognition},
  author={Feichtenhofer, Christoph and Fan, Haoqi and Malik, Jitendra and He, Kaiming},
  booktitle={Proceedings of the IEEE international conference on computer vision},
  pages={6202--6211},
  year={2019}
}

Model Zoo¶

Kinetics-400¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	inference_time(video/s)	gpu_mem(M)	ckpt	log	json
slowonly_r50_4x16x1_256e_kinetics400_rgb	short-side 256	8x4	ResNet50	None	72.76	90.51	x	3168	ckpt	log	json
slowonly_r50_video_4x16x1_256e_kinetics400_rgb	short-side 256	8	ResNet50	None	72.11	90.32	x	3168	ckpt	log	json
slowonly_r50_8x8x1_256e_kinetics400_rgb	short-side 256	8x4	ResNet50	None	74.42	91.49	x	5820	ckpt	log	json
slowonly_r50_4x16x1_256e_kinetics400_rgb	short-side 320	8x2	ResNet50	None	73.02	90.77	4.0 (40x3 frames)	3168	ckpt	log	json
slowonly_r50_8x8x1_256e_kinetics400_rgb	short-side 320	8x3	ResNet50	None	74.93	91.92	2.3 (80x3 frames)	5820	ckpt	log	json
slowonly_r50_4x16x1_256e_kinetics400_flow	short-side 320	8x2	ResNet50	ImageNet	61.79	83.62	x	8450	ckpt	log	json
slowonly_r50_8x8x1_196e_kinetics400_flow	short-side 320	8x4	ResNet50	ImageNet	65.76	86.25	x	8455	ckpt	log	json

Kinetics-400 Data Benchmark¶

In data benchmark, we compare two different data preprocessing methods: (1) Resize video to 340x256, (2) Resize the short edge of video to 320px, (3) Resize the short edge of video to 256px.

config	resolution	gpus	backbone	Input	pretrain	top1 acc	top5 acc	testing protocol	ckpt	log	json
slowonly_r50_randomresizedcrop_340x256_4x16x1_256e_kinetics400_rgb	340x256	8x2	ResNet50	4x16	None	71.61	90.05	10 clips x 3 crops	ckpt	log	json
slowonly_r50_randomresizedcrop_320p_4x16x1_256e_kinetics400_rgb	short-side 320	8x2	ResNet50	4x16	None	73.02	90.77	10 clips x 3 crops	ckpt	log	json
slowonly_r50_randomresizedcrop_256p_4x16x1_256e_kinetics400_rgb	short-side 256	8x4	ResNet50	4x16	None	72.76	90.51	10 clips x 3 crops	ckpt	log	json

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.

For more details on data preparation, you can refer to Kinetics400 in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train SlowOnly model on Kinetics-400 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/slowonly/slowonly_r50_4x16x1_256e_kinetics400_rgb.py \
    --work-dir work_dirs/slowonly_r50_4x16x1_256e_kinetics400_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test SlowOnly model on Kinetics-400 dataset and dump the result to a json file.

python tools/test.py configs/recognition/slowonly/slowonly_r50_4x16x1_256e_kinetics400_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json --average-clips=prob

For more details, you can refer to Test a dataset part in getting_started.

TIN¶

Introduction¶

@article{shao2020temporal,
    title={Temporal Interlacing Network},
    author={Hao Shao and Shengju Qian and Yu Liu},
    year={2020},
    journal={AAAI},
}

Model Zoo¶

Something-Something V1¶

|config | resolution | gpus | backbone| pretrain | top1 acc| top5 acc | reference top1 acc | reference top5 acc | gpu_mem(M) | ckpt | log| json| |:–|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:| |tin_r50_1x1x8_40e_sthv1_rgb|height 100|8x4| ResNet50 | ImageNet | 44.25 | 73.94 | 44.04 | 72.72 | 6181 | ckpt | log | json |

Something-Something V2¶

|config | resolution | gpus | backbone| pretrain | top1 acc| top5 acc | reference top1 acc | reference top5 acc | gpu_mem(M) | ckpt | log| json| |:–|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:| |tin_r50_1x1x8_40e_sthv2_rgb|height 240|8x4| ResNet50 | ImageNet | 56.38 | 83.53 | 56.38 | 83.50 | 6185 | ckpt | log | json |

Kinetics-400¶

|config | resolution | gpus | backbone| pretrain | top1 acc| top5 acc | gpu_mem(M) | ckpt | log| json| |:–|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:|:–:| |tin_tsm_finetune_r50_1x1x8_50e_kinetics400_rgb|short-side 256|8x4| ResNet50 | TSM-ImageNet | 70.89 | 89.89 | 6187 | ckpt | log | json |

Notes:

The reference topk acc are got by training the original repo ###1aacd0c with no AverageMeter issue. The AverageMeter issue will lead to incorrect performance, so we fix it before running.
The gpus indicates the number of gpu we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.
The values in columns named after “reference” are the results got by training on the original repo, using the same model settings.

For more details on data preparation, you can refer to Kinetics400, Something-Something V1 and Something-Something V2 in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train TIN model on Something-Something V1 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/tin/tin_r50_1x1x8_40e_sthv1_rgb.py \
    --work-dir work_dirs/tin_r50_1x1x8_40e_sthv1_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test TIN model on Something-Something V1 dataset and dump the result to a json file.

python tools/test.py configs/recognition/tin/tin_r50_1x1x8_40e_sthv1_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json

For more details, you can refer to Test a dataset part in getting_started.

TSM¶

Introduction¶

@inproceedings{lin2019tsm,
  title={TSM: Temporal Shift Module for Efficient Video Understanding},
  author={Lin, Ji and Gan, Chuang and Han, Song},
  booktitle={Proceedings of the IEEE International Conference on Computer Vision},
  year={2019}
}

@article{NonLocal2018,
  author =   {Xiaolong Wang and Ross Girshick and Abhinav Gupta and Kaiming He},
  title =    {Non-local Neural Networks},
  journal =  {CVPR},
  year =     {2018}
}

Model Zoo¶

Kinetics-400¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	reference top1 acc	reference top5 acc	inference_time(video/s)	gpu_mem(M)	ckpt	log	json
tsm_r50_1x1x8_50e_kinetics400_rgb	340x256	8	ResNet50	ImageNet	70.24	89.56	70.36	89.49	74.0 (8x1 frames)	7079	ckpt	log	json
tsm_r50_1x1x8_50e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	70.59	89.52	x	x	x	7079	ckpt	log	json
tsm_r50_video_1x1x8_50e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	70.25	89.66	70.36	89.49	74.0 (8x1 frames)	7077	ckpt	log	json
tsm_r50_dense_1x1x8_100e_kinetics400_rgb	340x256	8x4	ResNet50	ImageNet	72.9	90.44	72.22	90.37	11.5 (8x10 frames)	7079	ckpt	log	json
tsm_r50_dense_1x1x8_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	73.38	91.02	x	x	x	7079	ckpt	log	json
tsm_r50_1x1x16_50e_kinetics400_rgb	340x256	8	ResNet50	ImageNet	71.69	90.4	70.67	89.98	47.0 (16x1 frames)	10404	ckpt	log	json
tsm_r50_1x1x16_50e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	72.01	90.57	x	x	x	10398	ckpt	log	json
tsm_nl_embedded_gaussian_r50_1x1x8_50e_kinetics400_rgb	short-side 320	8x4	ResNet50	ImageNet	72.03	90.25	71.81	90.36	x	8931	ckpt	log	json
tsm_nl_gaussian_r50_1x1x8_50e_kinetics400_rgb	short-side 320	8x4	ResNet50	ImageNet	70.70	89.90	x	x	x	10125	ckpt	log	json
tsm_nl_dot_product_r50_1x1x8_50e_kinetics400_rgb	short-side 320	8x4	ResNet50	ImageNet	71.60	90.34	x	x	x	8358	ckpt	log	json

Something-Something V1¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	reference top1 acc	reference top5 acc	gpu_mem(M)	ckpt	log	json
tsm_r50_1x1x8_50e_sthv1_rgb	height 100	8	ResNet50	ImageNet	44.62	75.51	42.08	72.66	7077	ckpt	log	json

Something-Something V2¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	reference top1 acc	reference top5 acc	gpu_mem(M)	ckpt	log	json
tsm_r50_1x1x16_50e_sthv2_rgb	height 240	8	ResNet50	ImageNet	57.68	83.65	56.57	84.30	10400	ckpt	log	json
tsm_r101_1x1x8_50e_sthv2_rgb	height 240	8	ResNet101	ImageNet	59.12	85.74	59.20	85.27	9784	ckpt	log	json

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.
The values in columns named after “reference” are the results got by training on the original repo, using the same model settings.

For more details on data preparation, you can refer to Kinetics400, Something-Something V1 and Something-Something V2 in Data Preparation.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train TSM model on Kinetics-400 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/tsm/tsm_r50_1x1x8_50e_kinetics400_rgb.py \
    --work-dir work_dirs/tsm_r50_1x1x8_100e_kinetics400_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test TSM model on Kinetics-400 dataset and dump the result to a json file.

python tools/test.py configs/recognition/tsm/tsm_r50_1x1x8_50e_kinetics400_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json

For more details, you can refer to Test a dataset part in getting_started.

TSN¶

Introduction¶

@inproceedings{wang2016temporal,
  title={Temporal segment networks: Towards good practices for deep action recognition},
  author={Wang, Limin and Xiong, Yuanjun and Wang, Zhe and Qiao, Yu and Lin, Dahua and Tang, Xiaoou and Van Gool, Luc},
  booktitle={European conference on computer vision},
  pages={20--36},
  year={2016},
  organization={Springer}
}

Model Zoo¶

UCF-101¶

config	gpus	backbone	pretrain	top1 acc	top5 acc	gpu_mem(M)	ckpt	log	json
tsn_r50_1x1x3_80e_ucf101_rgb	8	ResNet50	ImageNet	80.12	96.09	8332	ckpt	log	json

Kinetics-400¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	reference top1 acc	reference top5 acc	inference_time(video/s)	gpu_mem(M)	ckpt	log	json
tsn_r50_1x1x3_100e_kinetics400_rgb	340x256	8	ResNet50	ImageNet	70.60	89.26	x	x	4.3 (25x10 frames)	8344	ckpt	log	json
tsn_r50_1x1x3_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	70.42	89.03	x	x	x	8343	ckpt	log	json
tsn_r50_dense_1x1x5_50e_kinetics400_rgb	340x256	8x3	ResNet50	ImageNet	70.18	89.10	69.15	88.56	12.7 (8x10 frames)	7028	ckpt	log	json
tsn_r50_320p_1x1x3_100e_kinetics400_rgb	short-side 320	8x2	ResNet50	ImageNet	70.91	89.51	x	x	10.7 (25x3 frames)	8344	ckpt	log	json
tsn_r50_320p_1x1x3_110e_kinetics400_flow	short-side 320	8x2	ResNet50	ImageNet	55.70	79.85	x	x	x	8471	ckpt	log	json
tsn_r50_320p_1x1x3_kinetics400_twostream [1: 1]*	x	x	ResNet50	ImageNet	72.76	90.52	x	x	x	x	x	x	x
tsn_r50_1x1x8_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	71.80	90.17	x	x	x	8343	ckpt	log	json
tsn_r50_320p_1x1x8_100e_kinetics400_rgb	short-side 320	8x3	ResNet50	ImageNet	72.41	90.55	x	x	11.1 (25x3 frames)	8344	ckpt	log	json
tsn_r50_320p_1x1x8_110e_kinetics400_flow	short-side 320	8x4	ResNet50	ImageNet	57.76	80.99	x	x	x	8473	ckpt	log	json
tsn_r50_320p_1x1x8_kinetics400_twostream [1: 1]*	x	x	ResNet50	ImageNet	74.64	91.77	x	x	x	x	x	x	x
tsn_r50_dense_1x1x8_100e_kinetics400_rgb	340x256	8	ResNet50	ImageNet	70.77	89.3	68.75	88.42	12.2 (8x10 frames)	8344	ckpt	log	json
tsn_r50_video_1x1x8_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	71.79	90.25	x	x	x	21558	ckpt	log	json
tsn_r50_video_dense_1x1x8_100e_kinetics400_rgb	short-side 256	8	ResNet50	ImageNet	70.4	89.12	x	x	x	21553	ckpt	log	json

Here, We use [1: 1] to indicate that we combine rgb and flow score with coefficients 1: 1 to get the two-stream prediction (without applying softmax).

Kinetics-400 Data Benchmark (8-gpus, ResNet50, ImageNet pretrain; 3 segments)¶

In data benchmark, we compare:

Different data preprocessing methods: (1) Resize video to 340x256, (2) Resize the short edge of video to 320px, (3) Resize the short edge of video to 256px;
Different data augmentation methods: (1) MultiScaleCrop, (2) RandomResizedCrop;
Different testing protocols: (1) 25 frames x 10 crops, (2) 25 frames x 3 crops.

config	resolution	training augmentation	testing protocol	top1 acc	top5 acc	ckpt	log	json
tsn_r50_multiscalecrop_340x256_1x1x3_100e_kinetics400_rgb	340x256	MultiScaleCrop	25x10 frames	70.60	89.26	ckpt	log	json
x	340x256	MultiScaleCrop	25x3 frames	70.52	89.39	x	x	x
tsn_r50_randomresizedcrop_340x256_1x1x3_100e_kinetics400_rgb	340x256	RandomResizedCrop	25x10 frames	70.11	89.01	ckpt	log	json
x	340x256	RandomResizedCrop	25x3 frames	69.95	89.02	x	x	x
tsn_r50_multiscalecrop_320p_1x1x3_100e_kinetics400_rgb	short-side 320	MultiScaleCrop	25x10 frames	70.32	89.25	ckpt	log	json
x	short-side 320	MultiScaleCrop	25x3 frames	70.54	89.39	x	x	x
tsn_r50_randomresizedcrop_320p_1x1x3_100e_kinetics400_rgb	short-side 320	RandomResizedCrop	25x10 frames	70.44	89.23	ckpt	log	json
x	short-side 320	RandomResizedCrop	25x3 frames	70.91	89.51	x	x	x
tsn_r50_multiscalecrop_256p_1x1x3_100e_kinetics400_rgb	short-side 256	MultiScaleCrop	25x10 frames	70.42	89.03	ckpt	log	json
x	short-side 256	MultiScaleCrop	25x3 frames	70.79	89.42	x	x	x
tsn_r50_randomresizedcrop_256p_1x1x3_100e_kinetics400_rgb	short-side 256	RandomResizedCrop	25x10 frames	69.80	89.06	ckpt	log	json
x	short-side 256	RandomResizedCrop	25x3 frames	70.48	89.89	x	x	x

Something-Something V1¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	reference top1 acc	reference top5 acc	gpu_mem(M)	ckpt	log	json
tsn_r50_1x1x8_50e_sthv1_rgb	height 100	8	ResNet50	ImageNet	18.55	44.80	17.53	44.29	10978	ckpt	log	json
tsn_r50_1x1x16_50e_sthv1_rgb	height 100	8	ResNet50	ImageNet	15.77	39.85	13.33	35.58	5691	ckpt	log	json

Something-Something V2¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	reference top1 acc	reference top5 acc	gpu_mem(M)	ckpt	log	json
tsn_r50_1x1x8_50e_sthv2_rgb	height 240	8x2	ResNet50	ImageNet	32.41	64.05	30.32	58.38	10978	ckpt	log	json
tsn_r50_1x1x16_50e_sthv2_rgb	height 240	8	ResNet50	ImageNet	22.48	49.08	22.50	47.29	5698	ckpt	log	json

Moments in Time¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	gpu_mem(M)	ckpt	log	json
tsn_r50_1x1x6_100e_mit_rgb	short-side 256	8x2	ResNet50	ImageNet	26.84	51.6	8339	ckpt	log	json

Multi-Moments in Time¶

config	resolution	gpus	backbone	pretrain	mAP	gpu_mem(M)	ckpt	log	json
tsn_r101_1x1x5_50e_mmit_rgb	short-side 256	8x2	ResNet101	ImageNet	61.09	10467	ckpt	log	json

ActivityNet v1.3¶

config	resolution	gpus	backbone	pretrain	top1 acc	top5 acc	gpu_mem(M)	ckpt	log	json
tsn_r50_320p_1x1x8_50e_activitynet_video_rgb	340x256	8x1	ResNet50	Kinetics400	73.97	93.46	5692	ckpt	log	json
tsn_r50_320p_1x1x8_50e_activitynet_clip_rgb	340x256	8x1	ResNet50	Kinetics400	76.07	94.10	5692	ckpt	log	json
tsn_r50_320p_1x1x8_150e_activitynet_video_flow	340x256	8x2	ResNet50	Kinetics400	58.70	84.72	5780	ckpt	log	json
tsn_r50_320p_1x1x8_150e_activitynet_clip_flow	340x256	8x2	ResNet50	Kinetics400	59.51	82.69	5780	ckpt	log	json

Notes:

The gpus indicates the number of gpu we used to get the checkpoint. It is noteworthy that the configs we provide are used for 8 gpus as default. According to the Linear Scaling Rule, you may set the learning rate proportional to the batch size if you use different GPUs or videos per GPU, e.g., lr=0.01 for 4 GPUs * 2 video/gpu and lr=0.08 for 16 GPUs * 4 video/gpu.
The inference_time is got by this benchmark script, where we use the sampling frames strategy of the test setting and only care about the model inference time, not including the IO time and pre-processing time. For each setting, we use 1 gpu and set batch size (videos per gpu) to 1 to calculate the inference time.
The values in columns named after “reference” are the results got by training on the original repo, using the same model settings.

For more details on data preparation, you can refer to preparing_ucf101, preparing_kinetics400, preparing_sthv1, preparing_sthv2, preparing_mit, preparing_mmit.

Train¶

You can use the following command to train a model.

python tools/train.py ${CONFIG_FILE} [optional arguments]

Example: train TSN model on Kinetics-400 dataset in a deterministic option with periodic validation.

python tools/train.py configs/recognition/tsn/tsn_r50_1x1x3_100e_kinetics400_rgb.py \
    --work-dir work_dirs/tsn_r50_1x1x3_100e_kinetics400_rgb \
    --validate --seed 0 --deterministic

For more details, you can refer to Training setting part in getting_started.

Test¶

You can use the following command to test a model.

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]

Example: test TSN model on Kinetics-400 dataset and dump the result to a json file.

python tools/test.py configs/recognition/tsn/tsn_r50_1x1x3_100e_kinetics400_rgb.py \
    checkpoints/SOME_CHECKPOINT.pth --eval top_k_accuracy mean_class_accuracy \
    --out result.json

For more details, you can refer to Test a dataset part in getting_started.