ValueError: Only call sparse_softmax_cross_entropy_with_logits with named arguments (labels=..., logits=..., ...)

[vidyaprashant]

Hi, whenever I am trying to train my own model which has only three .mrc images using the following query it results in some errors please anyone help me with solution

(tfg) C:\Users\admin\Desktop\my deeppicker>python train.py --train_type 1 --train_inputDir "mrc" --particle_size 50 --mrc_number 3 --particle_number 3 --coordinate_symbol "_manual_checked" --model_save_dir "trained_model" --model_save_file "model_demo_type1"
(0, 64, 64, 1) float64
(0,) int64
(0, 64, 64, 1) float64
(0,) int64
Load training data successfully!
Traceback (most recent call last):
File "train.py", line 176, in
main()
File "train.py", line 172, in main
train()
File "train.py", line 130, in train
deepModel.init_model_graph_train()
File "C:\Users\admin\Desktop\my deeppicker\deepModel.py", line 160, in init_model_graph_train
self.loss_operation = self.__loss(self.__inference(self.train_data_node,train=True))
File "C:\Users\admin\Desktop\my deeppicker\deepModel.py", line 91, in __loss
self.train_label_node, name = 'cross_entropy_all')
File "C:\Users\admin\Anaconda3\envs\tfg\lib\site-packages\tensorflow\python\ops\nn_ops.py", line 2011, in sparse_softmax_cross_entropy_with_logits
labels, logits)
File "C:\Users\admin\Anaconda3\envs\tfg\lib\site-packages\tensorflow\python\ops\nn_ops.py", line 1759, in _ensure_xent_args
"named arguments (labels=..., logits=..., ...)" % name)
ValueError: Only call sparse_softmax_cross_entropy_with_logits with named arguments (labels=..., logits=..., ...)

[nejyeah]

Hi, this is caused by the version of Tensorflow. This code is based on an old version of Tensorflow. The new version have changed some functions like the one "sparse_softmax_cross_entropy_with_logits"

[vidyaprashant]

will you please help me with the which version of Tensorflow is suitable for the existing code or which function needs to be used instead of "sparse_softmax_cross_entropy_with_logits"

[nejyeah]

The versions of Tensorflow suitable for this code is too old to get.
You may need to checkout the Tensorflow "sparse_softmax_cross_entropy_with_logits" API of the current version you use.