yu20103983 / FOTS

FOTS tensorflow implementation, Include train and test, EAST+Rotate+CRNN. FOTS: Fast Oriented Text Spotting with a Unified Network
89 stars 35 forks source link

TypeError when multigpu train #11

Open zhengjiawen opened 5 years ago

zhengjiawen commented 5 years ago

Traceback (most recent call last):
File "multigpu_train.py", line 221, in
tf.app.run()
File "/data/home/zjw/anaconda3/envs/tf-py27/lib/python2.7/site-packages/tensorflow/python/pla tform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "multigpu_train.py", line 113, in main
models.append(build_model(opt, reuse_variables, scope))
File "multigpu_train.py", line 76, in build_model
fots = FOTS_trainModel(input_images, input_brboxes, reuse_variables)
File "/data/home/zjw/pythonFile/FOTS-1/FOTS/fots_trainModel.py", line 22, in init
self.rg = recognizer.recognizer(self.rois, self.ws, NUM_CLASSES, FLAGS.keepProb, is_trainin g=True)
File "/data/home/zjw/pythonFile/FOTS-1/FOTS/recognizer.py", line 19, in init
self.convLayers(weight_decay, is_training)
File "/data/home/zjw/pythonFile/FOTS-1/FOTS/recognizer.py", line 63, in convLayers
self.word_vec = tf.transpose(tf.reshape(x, (-1, x.shape[2], x.shape[3])), (1, 0, 2))
File "/data/home/zjw/anaconda3/envs/tf-py27/lib/python2.7/site-packages/tensorflow/python/op /gen_array_ops.py", line 2619, in reshape
name=name)
File "/data/home/zjw/anaconda3/envs/tf-py27/lib/python2.7/site-packages/tensorflow/python/fra mework/op_def_library.py", line 493, in apply_op
raise err
TypeError: Failed to convert object of type <type 'tuple'> to Tensor. Contents: (-1, Dimension( 64), Dimension(128)). Consider casting elements to a supported type.

============================================================= The problem occurs when I train in ICDAR2015, can anyone help to fix it? Thanks a lot.

Marscep commented 5 years ago

do you fix it? I am struggling in it too.