[
  {
    "Top 1 Accuracy": 55.16,
    "code_links": [
      {
        "title": "swathikirans/GSM",
        "url": "https://github.com/swathikirans/GSM"
      }
    ],
    "date": "2019-12-01",
    "date2": 20191201,
    "model": "GSM Ensemble InceptionV3 (ImageNet pretrained)",
    "paper": {
      "title": "Gate-Shift Networks for Video Action Recognition",
      "url": "https://cknow.io/lib/54c2bb8752a1ab67"
    },
    "paper_data_uoa": "54c2bb8752a1ab67"
  },
  {
    "Top 1 Accuracy": 53.4,
    "code_links": [],
    "date": "2019-08-20",
    "date2": 20190820,
    "model": "GB + DF + LB (ResNet152, ImageNet pretrained)",
    "paper": {
      "title": "Action recognition with spatial-temporal discriminative filter banks",
      "url": "https://cknow.io/lib/ed09adc61dfdfb65"
    },
    "paper_data_uoa": "ed09adc61dfdfb65"
  },
  {
    "Top 1 Accuracy": 53,
    "code_links": [
      {
        "title": "craston/MARS",
        "url": "https://github.com/craston/MARS"
      }
    ],
    "date": "2019-06-01",
    "date2": 20190601,
    "model": "MARS+RGB+Flow (64 frames, Kinetics pretrained)",
    "paper": {
      "title": "MARS: Motion-Augmented RGB Stream for Action Recognition",
      "url": "https://cknow.io/lib/e917d94a27b04d45"
    },
    "paper_data_uoa": "e917d94a27b04d45"
  },
  {
    "Top 1 Accuracy": 51.68,
    "code_links": [
      {
        "title": "swathikirans/GSM",
        "url": "https://github.com/swathikirans/GSM"
      }
    ],
    "date": "2019-12-01",
    "date2": 20191201,
    "model": "GSM InceptionV3 (16 frames, ImageNet pretrained)",
    "paper": {
      "title": "Gate-Shift Networks for Video Action Recognition",
      "url": "https://cknow.io/lib/54c2bb8752a1ab67"
    },
    "paper_data_uoa": "54c2bb8752a1ab67"
  },
  {
    "Top 1 Accuracy": 50.7,
    "code_links": [],
    "date": "2019-08-07",
    "date2": 20190807,
    "model": "STM (16 frames, ImageNet pretraining)",
    "paper": {
      "title": "STM: SpatioTemporal and Motion Encoding for Action Recognition",
      "url": "https://cknow.io/lib/dbf65072610598a3"
    },
    "paper_data_uoa": "dbf65072610598a3"
  },
  {
    "Top 1 Accuracy": 50.7,
    "code_links": [
      {
        "title": "MIT-HAN-LAB/temporal-shift-module",
        "url": "https://github.com/MIT-HAN-LAB/temporal-shift-module"
      },
      {
        "title": "PaParaZz1/TemporalShiftModule",
        "url": "https://github.com/PaParaZz1/TemporalShiftModule"
      },
      {
        "title": "WavesUR/embedded_TSM",
        "url": "https://github.com/WavesUR/embedded_TSM"
      },
      {
        "title": "niveditarahurkar/CS231N-ActionRecognition",
        "url": "https://github.com/niveditarahurkar/CS231N-ActionRecognition"
      }
    ],
    "date": "2018-11-20",
    "date2": 20181120,
    "model": "TSM (RGB + Flow)",
    "paper": {
      "title": "TSM: Temporal Shift Module for Efficient Video Understanding",
      "url": "https://cknow.io/lib/0443f1df43248fc9"
    },
    "paper_data_uoa": "0443f1df43248fc9"
  },
  {
    "Top 1 Accuracy": 50,
    "code_links": [
      {
        "title": "zhoubolei/moments_models",
        "url": "https://github.com/zhoubolei/moments_models"
      },
      {
        "title": "metalbubble/moments_models",
        "url": "https://github.com/metalbubble/moments_models"
      },
      {
        "title": "thefonseca/predictive-coding",
        "url": "https://github.com/thefonseca/predictive-coding"
      },
      {
        "title": "shubhambitsg/activity-recognition",
        "url": "https://github.com/shubhambitsg/activity-recognition"
      }
    ],
    "date": "2018-01-09",
    "date2": 20180109,
    "model": "ResNet50 I3D (Moments pretrained)",
    "paper": {
      "title": "Moments in Time Dataset: one million videos for event understanding",
      "url": "https://cknow.io/lib/1224a44775a4689b"
    },
    "paper_data_uoa": "1224a44775a4689b"
  },
  {
    "Top 1 Accuracy": 49.7,
    "code_links": [],
    "date": "2019-08-27",
    "date2": 20190827,
    "model": "TRG-Inception-V3",
    "paper": {
      "title": "Temporal Reasoning Graph for Activity Recognition",
      "url": "https://cknow.io/lib/48eac55d1d0207b6"
    },
    "paper_data_uoa": "48eac55d1d0207b6"
  },
  {
    "Top 1 Accuracy": 49.5,
    "code_links": [],
    "date": "2019-08-27",
    "date2": 20190827,
    "model": "TRG-ResNet-50",
    "paper": {
      "title": "Temporal Reasoning Graph for Activity Recognition",
      "url": "https://cknow.io/lib/48eac55d1d0207b6"
    },
    "paper_data_uoa": "48eac55d1d0207b6"
  },
  {
    "Top 1 Accuracy": 49.3,
    "code_links": [
      {
        "title": "facebookresearch/VMZ",
        "url": "https://github.com/facebookresearch/VMZ"
      },
      {
        "title": "facebookresearch/R2Plus1D",
        "url": "https://github.com/facebookresearch/R2Plus1D"
      },
      {
        "title": "novav/CV_Spatiotemporal_Feature_3D_CNN",
        "url": "https://github.com/novav/CV_Spatiotemporal_Feature_3D_CNN"
      }
    ],
    "date": "2019-04-04",
    "date2": 20190404,
    "model": "ir-CSN-152",
    "paper": {
      "title": "Video Classification with Channel-Separated Convolutional Networks",
      "url": "https://cknow.io/lib/a9b1dc01cf004046"
    },
    "paper_data_uoa": "a9b1dc01cf004046"
  },
  {
    "Top 1 Accuracy": 49.2,
    "code_links": [
      {
        "title": "IuliaDuta/RSTG",
        "url": "https://github.com/IuliaDuta/RSTG"
      }
    ],
    "date": "2019-04-11",
    "date2": 20190411,
    "model": "RSTG (Kinetics pretrained)",
    "paper": {
      "title": "Recurrent Space-time Graph Neural Networks",
      "url": "https://cknow.io/lib/4454befce3a29b98"
    },
    "paper_data_uoa": "4454befce3a29b98"
  },
  {
    "Top 1 Accuracy": 48.6,
    "code_links": [
      {
        "title": "zhoubolei/moments_models",
        "url": "https://github.com/zhoubolei/moments_models"
      },
      {
        "title": "metalbubble/moments_models",
        "url": "https://github.com/metalbubble/moments_models"
      },
      {
        "title": "thefonseca/predictive-coding",
        "url": "https://github.com/thefonseca/predictive-coding"
      },
      {
        "title": "shubhambitsg/activity-recognition",
        "url": "https://github.com/shubhambitsg/activity-recognition"
      }
    ],
    "date": "2018-01-09",
    "date2": 20180109,
    "model": "ResNet50 I3D (Kinetics pretrained)",
    "paper": {
      "title": "Moments in Time Dataset: one million videos for event understanding",
      "url": "https://cknow.io/lib/1224a44775a4689b"
    },
    "paper_data_uoa": "1224a44775a4689b"
  },
  {
    "Top 1 Accuracy": 48.4,
    "code_links": [
      {
        "title": "facebookresearch/VMZ",
        "url": "https://github.com/facebookresearch/VMZ"
      },
      {
        "title": "facebookresearch/R2Plus1D",
        "url": "https://github.com/facebookresearch/R2Plus1D"
      },
      {
        "title": "novav/CV_Spatiotemporal_Feature_3D_CNN",
        "url": "https://github.com/novav/CV_Spatiotemporal_Feature_3D_CNN"
      }
    ],
    "date": "2019-04-04",
    "date2": 20190404,
    "model": "ir-CSN-101",
    "paper": {
      "title": "Video Classification with Channel-Separated Convolutional Networks",
      "url": "https://cknow.io/lib/a9b1dc01cf004046"
    },
    "paper_data_uoa": "a9b1dc01cf004046"
  },
  {
    "Top 1 Accuracy": 48.2,
    "code_links": [
      {
        "title": "s9xie/Mini-Kinetics-200",
        "url": "https://github.com/s9xie/Mini-Kinetics-200"
      },
      {
        "title": "kylemin/S3D",
        "url": "https://github.com/kylemin/S3D"
      }
    ],
    "date": "2017-12-13",
    "date2": 20171213,
    "model": "S3D-G (ImageNet pretrained)",
    "paper": {
      "title": "Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification",
      "url": "https://cknow.io/lib/54de448faaf57d21"
    },
    "paper_data_uoa": "54de448faaf57d21"
  },
  {
    "Top 1 Accuracy": 46.4,
    "code_links": [
      {
        "title": "mzolfaghari/ECO-efficient-video-understanding",
        "url": "https://github.com/mzolfaghari/ECO-efficient-video-understanding"
      }
    ],
    "date": "2018-04-24",
    "date2": 20180424,
    "model": "ECO-Net (ImageNet pretrained)",
    "paper": {
      "title": "ECO: Efficient Convolutional Network for Online Video Understanding",
      "url": "https://cknow.io/lib/4aaf0d4a7db158bc"
    },
    "paper_data_uoa": "4aaf0d4a7db158bc"
  },
  {
    "Top 1 Accuracy": 43.9,
    "code_links": [],
    "date": "2018-07-26",
    "date2": 20180726,
    "model": "Motion Feature Net",
    "paper": {
      "title": "Motion Feature Network: Fixed Motion Filter for Action Recognition",
      "url": "https://cknow.io/lib/60e476bb34d62b8a"
    },
    "paper_data_uoa": "60e476bb34d62b8a"
  },
  {
    "Top 1 Accuracy": 42.01,
    "code_links": [
      {
        "title": "metalbubble/TRN-pytorch",
        "url": "https://github.com/metalbubble/TRN-pytorch"
      },
      {
        "title": "okankop/MFF-pytorch",
        "url": "https://github.com/okankop/MFF-pytorch"
      }
    ],
    "date": "2017-11-22",
    "date2": 20171122,
    "model": "2-Stream TRN",
    "paper": {
      "title": "Temporal Relational Reasoning in Videos",
      "url": "https://cknow.io/lib/ab89498599fa31b3"
    },
    "paper_data_uoa": "ab89498599fa31b3"
  },
  {
    "Top 1 Accuracy": 41.97,
    "code_links": [],
    "date": "2019-05-29",
    "date2": 20190529,
    "model": "HF-TSN (ImageNet pretraining)",
    "paper": {
      "title": "Hierarchical Feature Aggregation Networks for Video Action Recognition",
      "url": "https://cknow.io/lib/dd1247b8980e18a8"
    },
    "paper_data_uoa": "dd1247b8980e18a8"
  },
  {
    "Top 1 Accuracy": 40.4,
    "code_links": [
      {
        "title": "craston/MARS",
        "url": "https://github.com/craston/MARS"
      }
    ],
    "date": "2019-06-01",
    "date2": 20190601,
    "model": "MARS+RGB+Flow (16 frames, Kinetics pretrained)",
    "paper": {
      "title": "MARS: Motion-Augmented RGB Stream for Action Recognition",
      "url": "https://cknow.io/lib/e917d94a27b04d45"
    },
    "paper_data_uoa": "e917d94a27b04d45"
  }
]