# global configs
Global:
  checkpoints: null
  output_dir: ./output/ 
  device: gpu
  save_interval: 20
  eval_during_train: True
  eval_interval: 1
  epochs: 100
  print_batch_step: 10
  use_visualdl: False
  # used for static mode and model export
  image_shape: [3, 224, 224]
  save_inference_dir: ./inference
  # training model under @to_static
  to_static: False


Arch:
  name: MoCo_finetune
  pretrained_model: ./pretrain/moco_v2_bs_256_epoch_200
  backbone:
    name: ResNet50
    stop_layer_name: avg_pool
    freeze_befor: avg_pool
  head:
    name: ClasHead
    class_num: 1000


# loss function config for traing/eval process
Loss:
  Train:
    - CELoss:
        weight: 1.0
  Eval:
    - CELoss:
        weight: 1.0


Optimizer:
  name: Momentum
  momentum: 0.9
  lr:
    name: MultiStepDecay
    learning_rate: 30.0
    milestones: [60, 80]


DataLoader:
  Train:
    dataset:
      name: MoCoImageNetDataset
      image_root: ./dataset/ILSVRC2012/
      cls_label_path: ./dataset/ILSVRC2012/train_list.txt
      return_label: True
      return_two_sample: False
      transform_ops:
        - DecodeImage:
            to_rgb: True
            channel_first: False
        - RandomResizedCrop:
            size: 224
        - RandomHorizontalFlip:
        - NormalizeImage:
            scale: 1.0/255.0
            mean: [0.485, 0.456, 0.406]
            std: [0.229, 0.224, 0.225]
            order: ''
    sampler:
      name: DistributedBatchSampler
      batch_size: 64
      shuffle: True
      drop_last: False
    loader:
      num_workers: 4
      use_shared_memory: True

  Eval:
    dataset:
      name: MoCoImageNetDataset
      image_root: ./dataset/ILSVRC2012/
      cls_label_path: ./dataset/ILSVRC2012/val_list.txt
      return_label: True
      return_two_sample: False
      transform_ops:
        - DecodeImage:
            to_rgb: True
            channel_first: False
        - Resize:
            size: 256
        - CenterCrop:
            size: 224
        - NormalizeImage:
            scale: 1.0/255.0
            mean: [0.485, 0.456, 0.406]
            std: [0.229, 0.224, 0.225]
            order: ''
    sampler:
      name: DistributedBatchSampler
      batch_size: 64
      shuffle: True
      drop_last: True
    loader:
      num_workers: 4
      use_shared_memory: True

Metric:
  Train:
    - TopkAcc:
        topk: [1, 5]
  Eval:
    - TopkAcc:
        topk: [1, 5]