ladatascience

pip uninstall <NOM DU PACKAGE>

# Par exemple, pour désinstaller numpy, on fera: pip uninstall numpy

ladatascience

import torch

is_free = torch.cuda.is_available()
if is_free:
  print("GPU disponible.")
else:
  print("Pas de GPU disponible.")

ladatascience

from skimage import data
from skimage.color import rgb2gray

# Chargement de l'image
img = rgb2gray(data.chelsea())

from scipy.ndimage import gaussian_filter

# Application d'un filtre gaussien
sigma = 5

ladatascience

import matplotlib.pyplot as plt
from skimage import data
from skimage.color import rgb2gray

img_color = data.chelsea()
plt.imshow(img_color)
plt.show()

img_gray = rgb2gray(img_color)
plt.imshow(img_gray, cmap="gray")

ladatascience

from skimage.color import rgb2gray

img_gray = rgb2gray(img_color) # où img_color est l'image couleur que vous souhaitez convertir en noir et blanc.

ladatascience

import skimage

FILE_NAME = "picture.jpg" # insérer le nom du fichier à charger.
img = skimage.io.imread(FILE_NAME)

ladatascience

import torch

free_GPU_mem, global_GPU_mem = torch.cuda.mem_get_info()
print(f"Free GPU memory: {free_GPU_mem/(1024 * 1024 * 1024)} GB")
print(f"Global GPU memory: {global_GPU_mem/(1024 * 1024 * 1024)} GB")