[SINGA-444] Can not run Model classes examples on Singa documentation

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Component/s: Documentation
• Environment: - python 3.6.8 ; - Ubuntu 18.10
• Assignee: Unassigned
• Reporter: thao p nguyen
• Created: 12/Apr/19 07:56
• Updated: 12/Apr/19 08:06
• Resolved: 12/Apr/19 08:06

Description

Following the Singa documentation, the API code for running models' example does not work. Below are messages:

1) FeedForward Net

>>> from singa import tensor
>>> from singa import loss
>>> x = tensor.Tensor((3, 5))
>>> x.uniform(0, 1)  # randomly genearte the prediction activation
>>> y = tensor.from_numpy(np.array([0, 1, 3], dtype=np.int))  # set the truth
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'np' is not defined
>>> f = loss.SoftmaxCrossEntropy()
>>> l = f.forward(True, x, y)  # l is tensor with 3 loss values
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'y' is not defined
>>> g = f.backward()  # g is a tensor containing all gradients of x w.r.t l
Segmentation fault (core dumped)

2) Loss

>>> from singa import tensor
>>> from singa import loss
>>>
>>> x = tensor.Tensor((3, 5))
>>> x.uniform(0, 1)  # randomly genearte the prediction activation
>>> y = tensor.from_numpy(np.array([0, 1, 3], dtype=np.int))  # set the truth
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'np' is not defined
>>>
>>> f = loss.SoftmaxCrossEntropy()
>>> l = f.forward(True, x, y)  # l is tensor with 3 loss values
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'y' is not defined
>>> g = f.backward()  # g is a tensor containing all gradients of x w.r.t l

3) >>> from singa import tensor
>>> from singa import metric
>>>
>>> x = tensor.Tensor((3, 5))
>>> x.uniform(0, 1) # randomly genearte the prediction activation
>>> x = tensor.SoftMax # normalize the prediction into probabilities
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: module 'singa.tensor' has no attribute 'SoftMax'
>>> y = tensor.from_numpy(np.array([0, 1, 3], dtype=np.int)) # set the truth
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
NameError: name 'np' is not defined
>>>
>>> f = metric.Accuracy()
>>> acc = f.evaluate(x, y) # averaged accuracy over all 3 samples in x