[深度学习][原创]使用labelImg+yolov5完成所有slowfast时空动作检测项目-训练测试篇_slowfast模型训练参数

当我们准备好训练集和配置文件后就可以着手训练的事情了。首先安装好slowfast环境，然后打开源码，slowfast/configs/AVA/SLOWFAST_32x2_R50_SHORT.yaml作如下配置

TRAIN:
  ENABLE: True
  DATASET: ava
  BATCH_SIZE: 4
  EVAL_PERIOD: 5
  CHECKPOINT_PERIOD: 1
  AUTO_RESUME: True
  # CHECKPOINT_FILE_PATH: /home/fut/Downloads/slowfast/pre-model/SLOWFAST_32x2_R101_50_50.pkl
  CHECKPOINT_TYPE: caffe2
DATA:
  NUM_FRAMES: 32
  SAMPLING_RATE: 2
  TRAIN_JITTER_SCALES: [256, 320]
  TRAIN_CROP_SIZE: 224
  TEST_CROP_SIZE: 224
  INPUT_CHANNEL_NUM: [3, 3]
  PATH_TO_DATA_DIR: './myava'
DETECTION:
  ENABLE: True
  ALIGNED: True
AVA:
  FRAME_DIR: 'myava/frame'
  FRAME_LIST_DIR: 'myava/frame_lists'
  ANNOTATION_DIR: 'myava/annotations'
  DETECTION_SCORE_THRESH: 0.8
  TRAIN_PREDICT_BOX_LISTS: [
    "ava_train_v2.2.csv",
    "person_box_67091280_iou90/ava_detection_train_boxes_and_labels_include_negative_v2.2.csv",
  ]
  TEST_PREDICT_BOX_LISTS: ["person_box_67091280_iou90/ava_detection_val_boxes_and_labels.csv"]
SLOWFAST:
  ALPHA: 4
  BETA_INV: 8
  FUSION_CONV_CHANNEL_RATIO: 2
  FUSION_KERNEL_SZ: 7
RESNET:
  ZERO_INIT_FINAL_BN: True
  WIDTH_PER_GROUP: 64
  NUM_GROUPS: 1
  DEPTH: 50
  TRANS_FUNC: bottleneck_transform
  STRIDE_1X1: False
  NUM_BLOCK_TEMP_KERNEL: [[3, 3], [4, 4], [6, 6], [3, 3]]
  SPATIAL_DILATIONS: [[1, 1], [1, 1], [1, 1], [2, 2]]
  SPATIAL_STRIDES: [[1, 1], [2, 2], [2, 2], [1, 1]]
NONLOCAL:
  LOCATION: [[[], []], [[], []], [[], []], [[], []]]
  GROUP: [[1, 1], [1, 1], [1, 1], [1, 1]]
  INSTANTIATION: dot_product
  POOL: [[[1, 2, 2], [1, 2, 2]], [[1, 2, 2], [1, 2, 2]], [[1, 2, 2], [1, 2, 2]], [[1, 2, 2], [1, 2, 2]]]
BN:
  USE_PRECISE_STATS: False
  NUM_BATCHES_PRECISE: 200
SOLVER:
  BASE_LR: 0.1
  LR_POLICY: steps_with_relative_lrs
  STEPS: [0, 10, 15, 20]
  LRS: [1, 0.1, 0.01, 0.001]
  MAX_EPOCH: 300
  MOMENTUM: 0.9
  WEIGHT_DECAY: 1e-7
  WARMUP_EPOCHS: 5.0
  WARMUP_START_LR: 0.000125
  OPTIMIZING_METHOD: sgd
MODEL:
  NUM_CLASSES: 2
  ARCH: slowfast
  MODEL_NAME: SlowFast
  LOSS_FUNC: bce
  DROPOUT_RATE: 0.5
  HEAD_ACT: sigmoid
TEST:
  ENABLE: True
  DATASET: ava
  BATCH_SIZE: 8
DATA_LOADER:
  NUM_WORKERS: 2
  PIN_MEMORY: True
NUM_GPUS: 4
NUM_SHARDS: 1
RNG_SEED: 0
OUTPUT_DIR: .

1 TRAIN:CHECKPOINT_FILE_PATH 就是我们下载的与训练模型的位置
2 DATA:PATH_TO_DATA_DIR 就是我们第二部分制作的数据集文件
3 AVA: 下面的路径也是对应第二部分数据集文件对应的地方
4 MODEL:NUM_CLASSES: 1 这里是最需要主义的，这里classes必需为2，因为我们有fight和person 2个分类。

开始训练：

 python tools/run_net.py --cfg configs/AVA/SLOWFAST_32x2_R50_SHORT5.yaml
训练出来的模型文件在chekpoints文件里面，我截图看看

惊喜不？是pyth格式文件，后面我们开始测试模型效果

首先 新建slowfast/demo/AVA/SLOWFAST_32x2_R50_SHORT.yaml文件，写下面代码

TRAIN:
  ENABLE: False
  DATASET: ava
  BATCH_SIZE: 1
  EVAL_PERIOD: 1
  CHECKPOINT_PERIOD: 1
  AUTO_RESUME: True
  CHECKPOINT_FILE_PATH: 'checkpoints/checkpoint_epoch_00140.pyth'  #path to pretrain model
  CHECKPOINT_TYPE: pytorch
DATA:
  NUM_FRAMES: 16
  SAMPLING_RATE: 2
  TRAIN_JITTER_SCALES: [256, 320]
  TRAIN_CROP_SIZE: 224
  TEST_CROP_SIZE: 256
  INPUT_CHANNEL_NUM: [3, 3]
DETECTION:
  ENABLE: True
  ALIGNED: False
AVA:
  BGR: False
  DETECTION_SCORE_THRESH: 0.8
  TEST_PREDICT_BOX_LISTS: ["person_box_67091280_iou90/ava_detection_val_boxes_and_labels.csv"]
SLOWFAST:
  ALPHA: 4
  BETA_INV: 8
  FUSION_CONV_CHANNEL_RATIO: 2
  FUSION_KERNEL_SZ: 5
RESNET:
  ZERO_INIT_FINAL_BN: True
  WIDTH_PER_GROUP: 64
  NUM_GROUPS: 1
  DEPTH: 101
  TRANS_FUNC: bottleneck_transform
  STRIDE_1X1: False
  NUM_BLOCK_TEMP_KERNEL: [[3, 3], [4, 4], [6, 6], [3, 3]]
  SPATIAL_DILATIONS: [[1, 1], [1, 1], [1, 1], [2, 2]]
  SPATIAL_STRIDES: [[1, 1], [2, 2], [2, 2], [1, 1]]
NONLOCAL:
  LOCATION: [[[], []], [[], []], [[6, 13, 20], []], [[], []]]
  GROUP: [[1, 1], [1, 1], [1, 1], [1, 1]]
  INSTANTIATION: dot_product
  POOL: [[[2, 2, 2], [2, 2, 2]], [[2, 2, 2], [2, 2, 2]], [[2, 2, 2], [2, 2, 2]], [[2, 2, 2], [2, 2, 2]]]
BN:
  USE_PRECISE_STATS: False
  NUM_BATCHES_PRECISE: 200
SOLVER:
  MOMENTUM: 0.9
  WEIGHT_DECAY: 1e-7
  OPTIMIZING_METHOD: sgd
MODEL:
  NUM_CLASSES: 2
  ARCH: slowfast
  MODEL_NAME: SlowFast
  LOSS_FUNC: bce
  DROPOUT_RATE: 0.5
  HEAD_ACT: sigmoid
TEST:
  ENABLE: False
  DATASET: ava
  BATCH_SIZE: 1
DATA_LOADER:
  NUM_WORKERS: 2
  PIN_MEMORY: True
 
NUM_GPUS: 1
NUM_SHARDS: 1
RNG_SEED: 0
OUTPUT_DIR: .
#TENSORBOARD:
#  MODEL_VIS:
#    TOPK: 2
DEMO:
  ENABLE: True
  LABEL_FILE_PATH: "myava/annotations/myava.json"
  INPUT_VIDEO: "myava/videos/fight1.mp4"
  OUTPUT_FILE: "myava/fight1_out.mp4"
  DETECTRON2_CFG: "COCO-Detection/faster_rcnn_R_50_FPN_3x.yaml"
  DETECTRON2_WEIGHTS: detectron2://COCO-Detection/faster_rcnn_R_50_FPN_3x/137849458/model_final_280758.pkl

由于我的显存6GB老是out of memory于是设置NUM_FRAMES为16，如果你显存够可以设置32，这个是默认值。

测试开始：

python tools/run_net.py --cfg demo/AVA/SLOWFAST_32x2_R50_SHORT.yaml

稍等一会结果就出来了。最终效果还可以。