Hi I wonder if there is an author can correct some mistakes in the document.
When I practice the example at Logistic regression explained — Apache MXNet documentation
at the paragrph(section 9), when I try to run the code ,
avg_train_loss = train_model() / train_data_size
I get a error like this
“raise TypeError('Operator {} registered in backend is known as {} in Python. ’
TypeError: Operator stack registered in backend is known as stack in Python. This is a legacy operator which can only accept legacy ndarrays, while received an MXNet numpy ndarray. Please call as_nd_ndarray() upon the numpy ndarray to convert it to a legacy ndarray, and then feed the converted array to this operator.”
probably its related to the codes that I changed, because when one(who is new to mxnet) try to run the example code, they will encouter the following error and must correct it
Section 1
mx.np.random.seed(12345) # Added for reproducibility
mx.np.random.seed(12345) # Added for reproducibility
otherwise, it will throw an error like np does not have function random()
Section 2
x = np.random.normal(0, 1, size=(size, 10), device=device)
x = np.random.normal(0, 1, size=(size, 10))
otherwise, it will tell the device does not work for normal function
Section 6
accuracy = mx.metric.Accuracy()
accuracy = mx.gluon.metric.Accuracy()
f1 = mx.metric.F1()
f1 = mx.gluon.metric.F1()
otherwise, it reports that mx.gluon does not have function of mertric.
I believe that some mistakes was caused by the early version and its not applied to version 1.7.0.
My question is, where I can contact someone to update the old codes?