modelscope/configs/examples/train.json

{
    "framework": "pytorch",

    "task": "image_classification",

    "model": {
        "type": "Resnet50ForImageClassification",
        "pretrained": null,
        "backbone": {
            "type": "ResNet",
            "depth": 50,
            "out_indices": [
                4
            ],
            "norm_cfg": {
                "type": "BN"
            }
        },
        "head": {
            "type": "ClsHead",
            "with_avg_pool": true,
            "in_channels": 2048,
            "loss_config": {
                "type": "CrossEntropyLossWithLabelSmooth",
                "label_smooth": 0
            },
            "num_classes": 1000
        }
    },

    "dataset": {
        "train": {
            "type": "ClsDataset",
            "data_source": {
                "list_file": "data/imagenet_raw/meta/train_labeled.txt",
                "root": "data/imagenet_raw/train/",
                "type": "ClsSourceImageList"
            }
        },
        "val": {
            "type": "ClsDataset",
            "data_source": {
                "list_file": "data/imagenet_raw/meta/val_labeled.txt",
                "root": "data/imagenet_raw/validation/",
                "type": "ClsSourceImageList"
            }
        }
    },


    "preprocessor":{
        "train": [
                {
                    "type": "RandomResizedCrop",
                    "size": 224
                },
                {
                    "type": "RandomHorizontalFlip"
                },
                {
                    "type": "ToTensor"
                },
                {
                    "type": "Normalize",
                    "mean": [
                        0.485,
                        0.456,
                        0.406
                    ],
                    "std": [
                        0.229,
                        0.224,
                        0.225
                    ]
                },
                {
                    "type": "Collect",
                    "keys": [
                        "img",
                        "gt_labels"
                    ]
                }
       ],
       "val": [
                {
                    "type": "Resize",
                    "size": 256
                },
                {
                    "type": "CenterCrop",
                    "size": 224
                },
                {
                    "type": "ToTensor"
                },
                {
                    "type": "Normalize",
                    "mean": [
                        0.485,
                        0.456,
                        0.406
                    ],
                    "std": [
                        0.229,
                        0.224,
                        0.225
                    ]
                },
                {
                    "type": "Collect",
                    "keys": [
                        "img",
                        "gt_labels"
                    ]
                }
       ]
    },

    "train": {
        "batch_size": 32,
        "learning_rate": 0.00001,
        "lr_scheduler_type": "cosine",
        "num_epochs": 20
    },

    "evaluation": {
        "batch_size": 32,
        "metrics": ["accuracy", "precision", "recall"]
    }

}