davrot/pytutorial
1
0
Fork 0
mirror of https://github.com/davrot/pytutorial.git synced 2025-04-16 12:16:42 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Update README.md

Browse source 
Signed-off-by: David Rotermund <54365609+davrot@users.noreply.github.com>
This commit is contained in:
David Rotermund 2023-12-16 19:34:58 +01:00 • committed by GitHub
parent 518a90e771
commit 8485d50177
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 54 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

66
pytorch/augmentation/README.md
View file

@ -280,6 +280,8 @@ plt.show()
### Auto Augment
#### CIFAR10
```python
random_auto1_transform = tv.transforms.AutoAugment(
tv.transforms.AutoAugmentPolicy.CIFAR10
@ -293,38 +295,78 @@ plt.show()
![image20](image20.png)
#### IMAGENET
```python
random_auto2_transform = tv.transforms.AutoAugment(
tv.transforms.AutoAugmentPolicy.IMAGENET
)
for i in range(1, 10):
new_image = random_auto2_transform((torch_image * 255).type(dtype=torch.uint8))
plt.subplot(3, 3, i)
plt.imshow(np.moveaxis(new_image.detach().numpy(), 0, 2))
plt.show()
```
![image](image.png)
![image21](image21.png)
#### SVHN
```python
random_auto3_transform = tv.transforms.AutoAugment(tv.transforms.AutoAugmentPolicy.SVHN)
for i in range(1, 10):
new_image = random_auto3_transform((torch_image * 255).type(dtype=torch.uint8))
plt.subplot(3, 3, i)
plt.imshow(np.moveaxis(new_image.detach().numpy(), 0, 2))
plt.show()
```
![image](image.png)
![image22](image22.png)
## Sequential
```python
sequential_transform = torch.nn.Sequential(
tv.transforms.RandomSolarize(threshold=0.5, p=1.0),
tv.transforms.RandomErasing(p=1.0),
)
new_image = sequential_transform((torch_image * 255).type(dtype=torch.uint8))
plt.imshow(np.moveaxis(new_image.detach().numpy(), 0, 2))
plt.show()
```
![image](image.png)
![image23](image23.png)
## Compose
```python
compose_transform = tv.transforms.Compose(
[
tv.transforms.RandomSolarize(threshold=0.5, p=1.0),
tv.transforms.RandomErasing(p=1.0),
]
)
new_image = compose_transform((torch_image * 255).type(dtype=torch.uint8))
plt.imshow(np.moveaxis(new_image.detach().numpy(), 0, 2))
plt.show()
```
![image](image.png)
![image24](image24.png)
# --------------
# Random Apply
```python
randomapply_transform = tv.transforms.RandomApply(
[
tv.transforms.RandomSolarize(threshold=0.5, p=1.0),
tv.transforms.RandomErasing(p=1.0),
],
p=0.5,
)
for i in range(1, 3):
plt.subplot(2, 1, i)
new_image = randomapply_transform((torch_image * 255).type(dtype=torch.uint8))
plt.imshow(np.moveaxis(new_image.detach().numpy(), 0, 2))
plt.show()
```
![image](image.png)
![image25](image25.png)