pytorch / vision

Datasets, Transforms and Models specific to Computer Vision

https://pytorch.org/vision

Repository from Github https://github.compytorch/vision

pytorch/vision Issues

Celeba dataset can not be downloaded
Updated 6 months ago2
Setting a list of `float` type values to `padding` argument of `RandomCrop()` gets the indirect error message
Closed 6 months ago1
Setting `float` and `bool` values to `quality` argument of `JPEG()` gets the indirect errors
Closed 6 months ago1
`num_ops` argument of `RandAugment()` shouldn't accept negative values
Closed 6 months ago1
Setting `5.3` and `9` to `bits` argument of `RandomPosterize()` gets an indirect error message
Closed 6 months ago1
Setting more than 2 elements to `scale` argument of `RandomResizedCrop()` works
Closed 7 months ago1
Setting more than 2 elements to `ratio` argument of `RandomResizedCrop()` works
Closed 7 months ago1
When we are going to get TorchVision CU128 for RTX 5000 series on Windows
Closed 7 months ago1
Setting the positive values `0`, `1`, `2`, ... and the negative values `-3`, `-2`, `-1`, ... to `magnitude` argument of `RandAugment()` works the same respectively
Closed 7 months ago1
Setting a float value to `magnitude` argument of `RandAugment()` gets the wrong error message
Updated 7 months ago1
Setting a float value to `num_magnitude_bins` argument of `RandAugment()` gets the indirect error message
Closed 7 months ago1
Torchvision Faster R-CNN onnx export with dynamic batch size fails during inference
Updated 7 months ago2
Setting the 1D list of 1 element to `RandomRotation()` and `RandomAffine()` works against an error message
Closed 7 months ago5
There should be the doc page for v2 `AutoAugmentPolicy`.
Closed 7 months ago1
`ConvertBoundingBoxFormat` should support `inplace`
Updated 7 months ago
Torchvision 0.20.1 looks for jpeg9 on MacOS, while depending on libjpeg-turbo which only provides jpeg8
Updated 7 months ago3
Some v2 transforms silently ignore numpy arrays.
Closed 8 months ago5
Different Behaviors of tranforms.ToTensor and transforms.v2.ToTensor
Closed 7 months ago1
Setting `-5` and `5` to `sharpness_factor` argument of `RandomAdjustSharpness()` sharpens and blurs images respectively
Updated 7 months ago1
Setting a list of one or two `float` values to `kernel_size` argument of `GaussianBlur()` gets an indirect error message
Closed 7 months ago1
For `kernel_size` parameter, the docs of `Conv1d()`, `Conv2d()` and `Conv3d()` don't explain which is width or height
Closed 7 months ago1
For `kernel_size` parameter, the doc of `GaussianBlur()` doesn't explain which is width or height
Updated 7 months ago1
Setting a complex value to `num_output_channels` argument of `Grayscale()` works
Closed 7 months ago1
`alpha` argument of `ElasticTransform()` should completely avoid negative values, giving error and the doc should have the explanation.
Updated 7 months ago1
the doc of `alpha` and `sigma` argument of `ElasticTransform()` should have the explanation about if it's a single value
Updated 7 months ago1
`sigma` argument of `ElasticTransform()` should completely avoid negative values, giving error and the doc should have the explanation.
Updated 7 months ago1
Setting a single negative value or a couple of negative values to `sigma` argument of `ElasticTransform()` doesn't get error against the error message
Updated 7 months ago1
Setting the 2D list of `str` type to `RandomRotation()`, `RandomHorizontalFlip()` and `RandomVerticalFlip()` after instantiation works
Closed 7 months ago2
Setting a 2D tensor to `RandomAffine()` after instantiation got other error message against error message
Updated 7 months ago1
Port vision to AIX
Closed 7 months ago2
Setting `0` and `1` to `p` argument of `RandomAutocontrast()` gets the same results
Updated 7 months ago1
The doc of `JPEG()` should say JPEG compression but not JPEG decompression
Closed 7 months ago1
[free threading] Support for free threading Python in torchvision
Updated 7 months ago
It's more intuitive that `sharpness_factor` argument of `RandomAdjustSharpness()` gives a blurred image with `x < 0`, gives an original image with `x = 0` and gives a sharpened image with `0 < x`.
Closed 7 months ago1
Can the `_make_divisible_function` be explained better?
Closed 7 months ago1
ONNX Model Runs into Runtime Error while Pytorch model is working fine
Closed 7 months ago1
torchvision_extra_decoders fails
Updated 7 months ago3
Different Behaviors of tranforms.Resize and transforms.v2.Resize
Closed 7 months ago1
Batched NMS generates onnx IF operator, which causes onnx conversion problem
Updated 7 months ago
LFW dataset support
Updated 7 months ago
Incorrect BoundingBoxes format conversion CXCYWH -> XYXY
Updated 8 months ago
SE module is missing in 'class FusedMBConv', 'efficientnet.py'. Is there a reason for it?
Closed 8 months ago5
[build] Linux aarch64 wheels are build with manylinux_2_34_aarch64 tag according to auditwheel check
Updated 8 months ago5
Order of offsets in deform_conv2d
Closed 8 months ago1
`size` parameter of `CenterCrop()` should work with `[width, height]` but not `[height, width]`
Closed 8 months ago1
initialized `RandomPerspective()` shows `p` parameter as the 1st
Closed 8 months ago1
`size` parameter of `Resize()` should work with `[width, height]` but not `[height, width]`
Closed 8 months ago1
torchvision version 0.14.0 with cuda version 116 support wheel file suddendly disappeard in download.pytorch.org
Closed 8 months ago
Setting `complex` values to `translate` argument of `RandomAffine()` gets indirect error messages
Closed 8 months ago1
Setting `complex` values to `degrees` argument of `RandomAffine()` gets indirect error messages
Closed 8 months ago1