def generate(nb, height, width,
max_nb_squares = 5, max_nb_properties = 10,
- many_colors = False):
+ nb_colors = 5):
- nb_colors = len(color_tokens) - 1 if many_colors else max_nb_squares
+ assert nb_colors >= max_nb_squares and nb_colors <= len(color_tokens) - 1
descr = [ ]
seen = {}
if len(d) != height * width: return []
+
for k, x in enumerate(d):
if x != color_names[0]:
if x in color_tokens:
return []
seen[x] = (color_id[x], k // width, k % width)
- square_c = torch.tensor( [ x[0] for x in seen.values() ] )
- square_i = torch.tensor( [ x[1] for x in seen.values() ] )
- square_j = torch.tensor( [ x[2] for x in seen.values() ] )
+ square_infos = tuple(zip(*seen.values()))
+ square_c = torch.tensor(square_infos[0])
+ square_i = torch.tensor(square_infos[1])
+ square_j = torch.tensor(square_infos[2])
s = all_properties(height, width, len(seen), square_i, square_j, square_c)
######################################################################
-def nb_missing_properties(descr, height, width):
+def nb_properties(descr, height, width):
if type(descr) == list:
- return [ nb_missing_properties(d, height, width) for d in descr ]
+ return [ nb_properties(d, height, width) for d in descr ]
d = descr.split('<img>', 1)
if len(d) == 0: return 0
descr = generate(nb = 5)
#print(descr2properties(descr))
- print(nb_missing_properties(descr))
+ print(nb_properties(descr))
with open('picoclvr_example.txt', 'w') as f:
for d in descr: