squeezenet1_1

torchvision.models.squeezenet1_1(*, weights: Optional[SqueezeNet1_1_Weights] = None, progress: bool = True, **kwargs: Any) → SqueezeNet

SqueezeNet 1.1 model from the official SqueezeNet repo.

SqueezeNet 1.1 has 2.4x less computation and slightly fewer parameters than SqueezeNet 1.0, without sacrificing accuracy.

Parameters:

• weights (SqueezeNet1_1_Weights, optional) – The pretrained weights to use. See SqueezeNet1_1_Weights below for more details, and possible values. By default, no pre-trained weights are used.

• progress (bool, optional) – If True, displays a progress bar of the download to stderr. Default is True.

• **kwargs – parameters passed to the torchvision.models.squeezenet.SqueezeNet base class. Please refer to the source code for more details about this class.

class torchvision.models.SqueezeNet1_1_Weights(value)

The model builder above accepts the following values as the weights parameter. SqueezeNet1_1_Weights.DEFAULT is equivalent to SqueezeNet1_1_Weights.IMAGENET1K_V1. You can also use strings, e.g. weights='DEFAULT' or weights='IMAGENET1K_V1'.

SqueezeNet1_1_Weights.IMAGENET1K_V1:

These weights reproduce closely the results of the paper using a simple training recipe. Also available as SqueezeNet1_1_Weights.DEFAULT.

acc@1 (on ImageNet-1K)	58.178
acc@5 (on ImageNet-1K)	80.624
categories	tench, goldfish, great white shark, … (997 omitted)
recipe	link
min_size	height=17, width=17
num_params	1235496
GFLOPS	0.35
File size	4.7 MB

The inference transforms are available at SqueezeNet1_1_Weights.IMAGENET1K_V1.transforms and perform the following preprocessing operations: Accepts PIL.Image, batched (B, C, H, W) and single (C, H, W) image torch.Tensor objects. The images are resized to resize_size=[256] using interpolation=InterpolationMode.BILINEAR, followed by a central crop of crop_size=[224]. Finally the values are first rescaled to [0.0, 1.0] and then normalized using mean=[0.485, 0.456, 0.406] and std=[0.229, 0.224, 0.225].