TypeError: Cannot handle this data type: (1, 1, 12), |u1

[VicZlq]

您好，我在训练的时候，出现这个问题 File "/home/anaconda3/envs/pt/lib/python3.7/site-packages/PIL/Image.py", line 2716, in fromarray raise TypeError("Cannot handle this data type: %s, %s" % typekey) TypeError: Cannot handle this data type: (1, 1, 12), |u1 我安装的PIL7.1.2的版本，是因为版本的问题么？ 请问有人遇到过同样的问题么？如何解决的呢？谢谢

[XueiuR]

我也遇到了同样的问题，环境是使用的google的colab平台 报错如下

/content/gdrive/My Drive/code/AttnGAN/code/miscc/utils.py:236: RuntimeWarning: invalid value encountered in true_divide
  one_map = (one_map - minV) / (maxV - minV)
Traceback (most recent call last):
  File "/usr/local/lib/python3.6/dist-packages/PIL/Image.py", line 2680, in fromarray
    mode, rawmode = _fromarray_typemap[typekey]
KeyError: ((1, 1, 12), '|u1')

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "main.py", line 146, in <module>
    gen_example(dataset.wordtoix, algo)  # generate images for customized captions
  File "main.py", line 83, in gen_example
    algo.gen_example(data_dic)
  File "/content/gdrive/My Drive/code/AttnGAN/code/trainer.py", line 514, in gen_example
    [attn_maps[j]], att_sze)
  File "/content/gdrive/My Drive/code/AttnGAN/code/miscc/utils.py", line 245, in build_super_images2
    PIL_att = Image.fromarray(np.uint8(one_map))
  File "/usr/local/lib/python3.6/dist-packages/PIL/Image.py", line 2682, in fromarray
    raise TypeError("Cannot handle this data type: %s, %s" % typekey)
TypeError: Cannot handle this data type: (1, 1, 12), |u1

请问你解决了吗？

[VicZlq]

您好，我在py3.7的版本下没有解决该问题，后来换到了2.7是可以正常运行的。

	赵柳

邮箱：18130039283@163.com |

签名由 网易邮箱大师 定制

On 06/26/2020 22:31, XueiuR wrote:

我也遇到了同样的问题，环境是使用的google的colab平台 报错如下

/content/gdrive/My Drive/code/AttnGAN/code/miscc/utils.py:236: RuntimeWarning: invalid value encountered in true_divide

one_map = (one_map - minV) / (maxV - minV)

Traceback (most recent call last):

File "/usr/local/lib/python3.6/dist-packages/PIL/Image.py", line 2680, in fromarray

mode, rawmode = _fromarray_typemap[typekey]

KeyError: ((1, 1, 12), '|u1')

During handling of the above exception, another exception occurred:

Traceback (most recent call last):

File "main.py", line 146, in

gen_example(dataset.wordtoix, algo)  # generate images for customized captions

File "main.py", line 83, in gen_example

algo.gen_example(data_dic)

File "/content/gdrive/My Drive/code/AttnGAN/code/trainer.py", line 514, in gen_example

[attn_maps[j]], att_sze)

File "/content/gdrive/My Drive/code/AttnGAN/code/miscc/utils.py", line 245, in build_super_images2

PIL_att = Image.fromarray(np.uint8(one_map))

File "/usr/local/lib/python3.6/dist-packages/PIL/Image.py", line 2682, in fromarray

raise TypeError("Cannot handle this data type: %s, %s" % typekey)

TypeError: Cannot handle this data type: (1, 1, 12), |u1

请问你解决了吗？

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[XueiuR]

好的 感谢！ 我的是python3.6，最后将25625612的向量截断为2562563了，然后可以运行，不知道会不会产生比较大的影响

[VicZlq]

好的！如果您跑出来了结果请告诉我一下修改方法！感谢！

	赵柳

邮箱：18130039283@163.com |

签名由 网易邮箱大师 定制

On 06/27/2020 01:59, XueiuR wrote:

好的 感谢！ 我的是python3.6，最后将25625612的向量截断为2562563了，然后可以运行，不知道会不会产生比较大的影响

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[XueiuR]

我将/code/miscc/util.py文件中，第243行的PIL_att = Image.fromarray(np.uint8(one_map))修改为PIL_att = Image.fromarray(np.uint8(one_map)[:,:,0:3])，第251行的row.append(np.concatenate([one_map, middle_pad], 1))修改为row.append(np.concatenate([one_map[:,:,0:3], middle_pad], 1))，然后重新运行，就可以生成图片结果了。 因为我才刚开始接触GAN和CV，所以我不确定这种修改方式会不会对生成的结果产生什么影响。

[XueiuR]

我刚才也试了使用python2.7运行，效果会比经过我修改后使用python3生成的图片，从直观效果上看会好很多

[Easonshow]

您好，我最近也在研究attngan 也遇到了这个问题一直不知道怎么解决？方便加个好友吗 ，相互学习一下，QQ 1014976349

[hs457681503]

我刚才也试了使用python2.7运行，效果会比经过我修改后使用python3生成的图片，从直观效果上看会好很多

你好，我也出现了这个问题，请问你是怎么改的代码呢？可以加我吗？qq是 457681503

[savitha91]

Hi, regarding the issue, replace the line one_map = skimage.transform.pyramid_expand(one_map, sigma=20, upscale=vis_size // att_sze) with this one_map = skimage.transform.pyramid_expand(one_map, sigma=20, upscale=vis_size // att_sze, multichannel=True) inside the method utils.build_super_images2()