Create a model Args: model_name (str): name of model to instantiate pretrained (bool): load pretrained ImageNet-1k weights if tru

See http://www.codinghorror.com/blog/archives/001018.html

Selective Kernel Attention Module Selective Kernel attention mechanism factored out into its own module.

At least `x` of `a` and `b` `Tensors` are equal.

Make a random crop of image_size.

Decode block definition string Gets a list of block arg (dicts) through a string notation of arguments. E.g. ir_r2_k3_s2_e1_i32_o16_se0.25_n

Random horizontal image flip.

FBNet-C Paper: https://arxiv.org/abs/1812.03443 Ref Impl: https://github.com/facebookresearch/maskrcnn-benchmark/blob/master/maskrcn

Creates a mnasnet-b1 model. Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet Paper: https://arxiv.org/pdf/1807

Generate MobileNet-V2 network Ref impl: https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py Paper:

Creates a MobileNet-V3 model. Ref impl: ? Paper: https://arxiv.org/abs/1905.02244 Args: channel_multiplier: multiplier to number o

Creates the Single-Path NAS model from search targeted for Pixel1 phone. Paper: https://arxiv.org/abs/1904.02877 Args: channel_multipl

Per-stage depth scaling Scales the block repeats in each stage. This depth scaling impl maintains compatibility with the EfficientNet scaling

Create AugMix PyTorch transform :param config_str: String defining configuration of random augmentation. Consists of multiple sections separated

Create a AutoAugment transform :param config_str: String defining configuration of auto augmentation. Consists of multiple sections separate

Recursively traverse module and its children to replace all instances of ``torch.nn.modules.batchnorm._BatchNorm`` with `SplitBatchnorm2d`.

Generates cropped_image using one of the bboxes randomly distorted. See `tf.image.sample_distorted_bounding_box` for more documentation. Arg