3 This is the port of the Synthetic Visual Reasoning Test to the pytorch
9 torch.ByteTensor svrt.generate_vignettes(int problem_number, torch.LongTensor labels)
14 * `problem_number` indicates which of the 23 problem to use
15 * `labels` indicates the boolean labels of the vignettes to generate
17 The returned ByteTensor has three dimensions:
23 The two additional functions
26 torch.ByteStorage svrt.compress(torch.ByteStorage x)
32 torch.ByteStorage svrt.uncompress(torch.ByteStorage x)
35 provide a lossless compression scheme adapted to the ByteStorage of
36 the vignette ByteTensor (i.e. expecting a lot of 255s, a few 0s, and
39 They allow to reduce the memory footprint by a factor ~50, and may be
40 usefull to deal with very large data-sets and avoid re-generating
41 images at every batch.
43 See vignette_set.py for a class CompressedVignetteSet using it.
45 # Installation and test #
54 should generate an image example.png in the current directory.
56 Note that the image generation does not take advantage of GPUs or
57 multi-core, and can be as fast as 10,000 vignettes per second and as
58 slow as 40 on a 4GHz i7-6700K.