[問題] np.stack 錯誤問題

[gigigigi (gigigigi)]

https://gist.github.com/shihyu/a56fb45dfb7781a48bcb7199cad25a0f


下面寫成一行

train_data = np.stack([np.array(Image.open("./data/56_imitate_train_set/" +
'{0:05}'.format(int(row[0])) + ".jpg"))/255.0 for row in
csv.reader(traincsv)])


會出現 MemoryError, 下面錯誤訊息

Traceback (most recent call last):
  File "test.py", line 22, in <module>
    train_data = np.stack([np.array(Image.open("./data/56_imitate_train_set/"
+ '{0:05}'.format(int(row[0])) + ".jpg"))/255.0 for row in
csv.reader(traincsv)])
  File
"/home/shihyu/anaconda3/lib/python3.6/site-packages/numpy/core/shape_base.py",
line 360, in stack
    return _nx.concatenate(expanded_arrays, axis=axis, out=out)
MemoryError




但是把一行下面拆開來寫是ok

for row in csv.reader(traincsv):
    tmp = [np.array(Image.open("./data/56_imitate_train_set/" +
'{0:05}'.format(int(row[0])) + ".jpg"))/255.0]

train_data = np.stack(tmp)

請問這是什麼原因？

謝謝

--
※ 發信站: 批踢踢實業坊(ptt.cc), 來自: 175.181.180.179
※ 文章網址: https://www.ptt.cc/bbs/Python/M.1529143399.A.F63.html
推 brahmsliu: 你把tmp 在每次回圈都重新定義了？所以不會爆