squeezenet1_0¶
- torchvision.models.squeezenet1_0(*, weights: Optional[SqueezeNet1_0_Weights] = None, progress: bool = True, **kwargs: Any) SqueezeNet [source]¶
SqueezeNet model architecture from the SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size paper.
- Parameters:
weights (
SqueezeNet1_0_Weights
, optional) – The pretrained weights to use. SeeSqueezeNet1_0_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_0_Weights(value)[source]¶
The model builder above accepts the following values as the
weights
parameter.SqueezeNet1_0_Weights.DEFAULT
is equivalent toSqueezeNet1_0_Weights.IMAGENET1K_V1
. You can also use strings, e.g.weights='DEFAULT'
orweights='IMAGENET1K_V1'
.SqueezeNet1_0_Weights.IMAGENET1K_V1:
These weights reproduce closely the results of the paper using a simple training recipe. Also available as
SqueezeNet1_0_Weights.DEFAULT
.acc@1 (on ImageNet-1K)
58.092
acc@5 (on ImageNet-1K)
80.42
categories
tench, goldfish, great white shark, … (997 omitted)
recipe
min_size
height=21, width=21
num_params
1248424
The inference transforms are available at
SqueezeNet1_0_Weights.IMAGENET1K_V1.transforms
and perform the following preprocessing operations: AcceptsPIL.Image
, batched(B, C, H, W)
and single(C, H, W)
imagetorch.Tensor
objects. The images are resized toresize_size=[256]
usinginterpolation=InterpolationMode.BILINEAR
, followed by a central crop ofcrop_size=[224]
. Finally the values are first rescaled to[0.0, 1.0]
and then normalized usingmean=[0.485, 0.456, 0.406]
andstd=[0.229, 0.224, 0.225]
.