ResNet152_vd.yaml 1.1 KB

1234567891011121314151617181920212223242526272829303132333435363738394041
  1. Global:
  2. model: ResNet152_vd
  3. mode: check_dataset # check_dataset/train/evaluate/predict
  4. dataset_dir: "/paddle/dataset/paddlex/cls/cls_flowers_examples"
  5. device: gpu:0,1,2,3
  6. output: "output"
  7. CheckDataset:
  8. convert:
  9. enable: False
  10. src_dataset_type: null
  11. split:
  12. enable: False
  13. train_percent: null
  14. val_percent: null
  15. Train:
  16. num_classes: 102
  17. epochs_iters: 20
  18. batch_size: 64
  19. learning_rate: 0.1
  20. pretrain_weight_path: https://paddle-model-ecology.bj.bcebos.com/paddlex/official_pretrained_model/ResNet152_vd_pretrained.pdparams
  21. warmup_steps: 5
  22. resume_path: null
  23. log_interval: 1
  24. eval_interval: 1
  25. save_interval: 1
  26. Evaluate:
  27. weight_path: "output/best_model/best_model.pdparams"
  28. log_interval: 1
  29. Export:
  30. weight_path: https://paddle-model-ecology.bj.bcebos.com/paddlex/official_pretrained_model/ResNet152_vd_pretrained.pdparams
  31. Predict:
  32. batch_size: 1
  33. model_dir: "output/best_model/inference"
  34. input: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_image_classification_001.jpg"
  35. kernel_option:
  36. run_mode: paddle