Generates per block widths from RegNet parameters.

Get parameters for ``crop`` for a random sized crop. Args: img (PIL Image): Image to be cropped. scale (tuple): range

Check if a model name exists

Check if a model exists within a subset of modules Args: model_name (str) - name of model to check module_names (tuple, list, set)

r"""Loads a custom (read non .pth) weight file Downloads checkpoint file into cache-dir like torch.hub based loaders, but calls a passed in c

Load pretrained checkpoint Args: model (nn.Module) : PyTorch model module default_cfg (Optional[Dict]): default configuration fo

MNASNet B1, depth multiplier of 1.0.

Fetch a model entrypoint for specified model name

feature module name accessor

cfg/argparse to kwargs helper Convert optimizer args in argparse args or cfg like object to keyword args for updated create fn.

Preprocesses the given image for evaluation. Args: image_bytes: `Tensor` representing an image binary of arbitrary size. use_bfloat16

Preprocesses the given image for evaluation. Args: image_bytes: `Tensor` representing an image binary of arbitrary size. use_bfloat16

Preprocesses the given image. Args: image_bytes: `Tensor` representing an image binary of arbitrary size. is_training: `bool` for whe

Converts a float to closest non-zero int divisible by q.

Create a RandAugment transform :param config_str: String defining configuration of random augmentation. Consists of multiple sections separa

Standard CutMix bounding-box Generates a random square bbox based on lambda value. This impl includes support for enforcing a border margin a

Min-Max CutMix bounding-box Inspired by Darknet cutmix impl, generates a random rectangular bbox based on min/max percent values applied to e

feature reduction (output stride) accessor

MNASNet A1 (w/ SE), depth multiplier of 1.0.

Constructs a SE-ResNet-26-T model. This is technically a 28 layer ResNet, like a 'D' bag-of-tricks model but with tiered 24, 32, 64 channels i

Computes the Hutchinson approximation of the hessian trace and accumulates it for each trainable parameter.

If separate==True, the transforms are returned as a tuple of 3 separate transforms for use in a mixing dataset that passes * all data th

Update the default_cfg and kwargs before passing to model FIXME this sequence of overlay default_cfg, set default kwargs, filter kwargs coul

R50+ViT-B/16 hybrid model from original paper (https://arxiv.org/abs/2010.11929). ImageNet-21k weights @ 224x224, source https://github.com/googl

R50+ViT-B/16 hybrid from original paper (https://arxiv.org/abs/2010.11929). ImageNet-1k weights fine-tuned from in21k @ 384x384, source https://g

Args: windows: (num_windows*B, window_size, window_size, C) window_size (int): Window size H (int): Height of image

Zeros out the accumalated hessian traces.

If `module` is IABN don't use half precision.

Args: img (PIL Image): Image to be cropped and resized. Returns: PIL Image: Randomly cropped and resized ima