konverner

import torch
from torchvision import models


modules=list(models.resnet18(weights='IMAGENET1K_V1').children())[:-1]
model = nn.Sequential(*modules)

x = torch.rand(size=[1, 3, 640, 480])
emb = model(x)

konverner

def grad_norm(model: nn.Module):
  total_norm = 0
  for p in model.parameters():
      param_norm = p.grad.detach().data.norm(2)
      total_norm += param_norm.item() ** 2
  return total_norm ** 0.5

konverner

def mixup(X : torch.tensor,
          Y: torch.tensor,
          alpha: int=8,
          device: str='cpu'):
  """
  it performs mixup between the first half of the batch with the second one

  params
  ---
  X : images tensor [bs, c, h, w]

konverner

column = ...
cond1, cond2 = ..., ...
agg_func = 'mean'

if agg_func == 'mean':
  new_value = df[(cond1) & (cond2)][column].mean()
else:
  new_value = df[(cond1) & (cond2)][column].mode()[0]
  
df.loc[(cond1) & (cond2), column] = df.loc[(cond1) & (cond2)][column].fillna(new_value)

konverner

X_train = ... # (n,2) numpy array dataset

# Create a meshgrid
x_min, x_max = X_train[:, 0].min() - 1, X_train[:, 0].max() + 1
y_min, y_max = X_train[:, 1].min() - 1, X_train[:, 1].max() + 1
xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.1),
                     np.arange(y_min, y_max, 0.1))
                     
                     
# Make predictions on the meshgrid

konverner

pd.options.display.float_format = '{:,.1f}%'.format
print((df.isna().sum()/df.count()).sort_values())

konverner

N_before = len(df)
COLUMNS_TO_PROCESS = ['sqft_living', 'sqft_basement', 'bedrooms']
TRESHOLD_Q3 = 0.95
TRESHOLD_Q1 = 0.05

for column_name in COLUMNS_TO_PROCESS:
    Q3 = df[column_name].quantile(TRESHOLD_Q3)
    Q1 = df[column_name].quantile(TRESHOLD_Q1)
    print(f"{column_name} [Q1, Q3] : [{Q1}, {Q3}]")
    IQR = Q3 - Q1

konverner

from PIL import Image
import requests
from io import BytesIO


url = 'https://upload.wikimedia.org/wikipedia/en/7/7d/Lenna_%28test_image%29.png'
response = requests.get(url)
img = Image.open(BytesIO(response.content))

konverner

from pathlib import Path

DIR = Path.cwd() # work directory
PATH_TEST_DIR = Path(DIR, 'data')
PATH_TEST_FILE = Path( DIR, 'file.csv')

konverner

import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader
from torchvision.datasets import MNIST
from torchvision.transforms import ToTensor
from tqdm import tqdm

# Set device (GPU or CPU)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")