use PIL's point-function to accelerate adjust_gamma (#410)

torchvision/transforms/functional.py

@@ -523,11 +523,10 @@ def adjust_gamma(img, gamma, gain=1):
     input_mode = img.mode
     img = img.convert('RGB')
 
-    np_img = np.array(img, dtype=np.float32)
-    np_img = 255 * gain * ((np_img / 255) ** gamma)
-    np_img = np.uint8(np.clip(np_img, 0, 255))
+    gamma_map = [255 * gain * pow(ele / 255., gamma) for ele in range(256)] * 3
+    img = img.point(gamma_map)  # use PIL's point-function to accelerate this part
 
-    img = Image.fromarray(np_img, 'RGB').convert(input_mode)
+    img = img.convert(input_mode)
     return img