search

komiya-m / MirrorGAN

27 stars 12 forks source link

No model found #1

Open Jillian555 opened 5 years ago

Jillian555 commented 5 years ago

OSError: Unable to open file (unable to open file: name = 'model/cnn_rnn_encoder02_2.h5', errno = 2, error message = 'No such file or directory', flags = 0, o_flags = 0)

ymmtr6 commented 5 years ago

I think that is an error at the pre-training.

before run pre-train,

$ mkdir model

and change config.py 33 : C.TRAIN.RNN_DEC = 'model/cnn_rnn_encoder02_2.h5' ↓ 33 : C.TRAIN.RNN_DEC = ""

nana33 commented 5 years ago

I think that is an error at the pre-training.

before run pre-train,

$ mkdir model

and change config.py 33 : C.TRAIN.RNN_DEC = 'model/cnn_rnn_encoder02_2.h5' ↓ 33 : C.TRAIN.RNN_DEC = ""

hello，i change config.py and get

Total params: 30,012,522 Trainable params: 8,209,738 Non-trainable params: 21,802,784

but File "E:\data\MirrorGAN-master\generator.py", line 106, in next img256_ar = np.empty((0, self.imsize[2], self.imsize[2], 3))

IndexError: list index out of range new error happen i want to know whether you have the same problem, thank you

ymmtr6 commented 5 years ago

sorry, there was another change in "config.py" to pre-train.

20: __C.TREE.BRANCH_NUM = 1
↓
20: __C.TREE.BRANCH_NUM = 3

This is because the number of elements in the "self.imsize" array is changed by " __C.TREE.BRANCH_NUM". Probably ...

mengistuabebe06 commented 3 years ago

sorry, there was another change in "config.py" to pre-train.

20: __C.TREE.BRANCH_NUM = 1
↓
20: __C.TREE.BRANCH_NUM = 3

This is because the number of elements in the "self.imsize" array is changed by " __C.TREE.BRANCH_NUM". Probably ...

I think that is an error at the pre-training.

before run pre-train,

$ mkdir model

and change config.py 33 : C.TRAIN.RNN_DEC = 'model/cnn_rnn_encoder02_2.h5' ↓ 33 : C.TRAIN.RNN_DEC = ""

does it work ?