Table of Contents:
import segmentation_models_pytorch as smp
model = smp.from_pretrained("<save-directory-or-this-repo>")
model_init_params = {
"encoder_name": "se_resnext101_32x4d",
"encoder_depth": 5,
"encoder_weights": "imagenet",
"decoder_use_norm": "batchnorm",
"decoder_channels": (256, 128, 64, 32, 16),
"decoder_attention_type": None,
"decoder_interpolation": "nearest",
"in_channels": 3,
"classes": 1,
"activation": None,
"aux_params": None
}
[
{
"test_per_image_iou": 0.1955837905406952,
"test_dataset_iou": 0.14369702339172363,
"test_per_image_accuracy": 0.9666634202003479,
"test_dataset_accuracy": 0.9666634202003479
}
]
Dataset name: VIP
This model has been pushed to the Hub using the PytorchModelHubMixin