jd icon indicating copy to clipboard operation
jd copied to clipboard

用python2.7 32位编译train.py出现EOFError怎么解决

Open huyp182 opened this issue 7 years ago • 7 comments

求解决..

huyp182 avatar May 15 '17 10:05 huyp182

你是不是用了 pkl了? 删除那个pkl 就可以了

ZS167275 avatar May 15 '17 13:05 ZS167275

用了numpy+mkl。 不太清楚pkl指的是什么,但是应该没用这个吧。 这是cmd里的: C:\Users\Administrator>pip uninstall packle Cannot uninstall requirement packle, not installed

C:\Users\Administrator>pip uninstall pkl Cannot uninstall requirement pkl, not installed

huyp182 avatar May 16 '17 07:05 huyp182

你说的pkl是pickle.py啊...才发现,我试试

huyp182 avatar May 16 '17 07:05 huyp182

我是说删除它的中间结果就行 就是产生的中间结果

ZS167275 avatar May 16 '17 11:05 ZS167275

谢谢,那个问题解决了,又出现了新问题.. Traceback (most recent call last): File "train.py", line 82, in xgboost_make_submission() File "train.py", line 18, in xgboost_make_submission user_index, training_data, label = make_train_set(train_start_date, train_end_date, test_start_date, test_end_date) File "D:\jd-master\gen_feat.py", line 272, in make_train_set user = get_basic_user_feat() File "D:\jd-master\gen_feat.py", line 46, in get_basic_user_feat user = pickle.load(open(dump_path)) TypeError: a bytes-like object is required, not 'str'

把dump_path改成dump_path.encode("utf-8")也没用,一样的错误

huyp182 avatar May 17 '17 05:05 huyp182

user = get_basic_user_feat() 你看看这个函数的问题

ZS167275 avatar May 17 '17 07:05 ZS167275