search

ruotianluo / DiscCaptioning

Code for Discriminability objective for training descriptive captions(CVPR 2018)
110 stars 21 forks source link

IOError: Failed to interpret file 'data/cocotalk_fc/100000.npy' as a pickle. #5

Closed dongqiaqia closed 6 years ago

dongqiaqia commented 6 years ago

Hi, ruotian. When I ran ‘bash run_fc_con.sh’, the training is terminated by the above error. I have checked the '100000.npy ', which is 1.8G and cannot be loaded by numpy. So I wonder if this file is not right.

ruotianluo commented 6 years ago

Did you download the preextracted features and put them in the right space!?

dongqiaqia commented 6 years ago

I am appreciated for your quick answer. I have downloaded the preextracted features from the link given by you, and put the fc features into 'data/cocotalk_fc'. The training was running several iterations and terminated by the obove error.

ruotianluo commented 6 years ago

Sorry, I didn't read it carefully. That looks weird to me. I will try to get back tomorrow

dongqiaqia commented 6 years ago

Never mind. It' s kind of you to help.

ruotianluo commented 6 years ago

There's definitely something went wrong.

ruotianluo commented 6 years ago

Try the current one?

dongqiaqia commented 6 years ago

The error is solved. Thanks!!