Incorrect Conversion from md&ipynb
LinkHS opened this issue · 0 comments
1. symbol.md
The contents in symbol.md (see attachment symbol.md
) is like below. When I run "notedown symbol.md --to markdown --strip" (see chapter 1 and attachment symbol.md), then it becomes like chapter 2.
ex = c.eval(ctx = mx.cpu(), a = mx.nd.ones([2,3]), b = mx.nd.ones([2,3]))
print('number of outputs = %d\nthe first output = \n%s' % (
len(ex), ex[0].asnumpy()))
For neural nets, a more commonly used pattern is simple_bind
, which
creates all of the argument arrays for you. Then you can call forward
,
and backward
(if the gradient is needed) to get the gradient.
Load and Save
Logically symbols correspond to ndarrays. They both represent a tensor. They both
are inputs/outputs of operators. We can either serialize a Symbol
object by
using pickle
, or by using save
and load
methods directly as we discussed in
NDArray tutorial.
2. symbol_result.md
ex = c.eval(ctx = mx.cpu(), a = mx.nd.ones([2,3]), b = mx.nd.ones([2,3]))
print('number of outputs = %d\nthe first output = \n%s' % (
len(ex), ex[0].asnumpy()))
For neural nets, a more commonly used pattern is simple_bind
, which
creates all of the argument arrays for you. Then you can call forward
,
and
backward
(if the gradient is needed) to get the gradient.
2. Load and
Save
Logically symbols correspond to ndarrays. They both represent a tensor.
They both
are inputs/outputs of operators. We can either serialize a Symbol
object by
using pickle
, or by using save
and load
methods directly as we
discussed in
[NDArray tutorial](http://mxnet.io/tutorials/basic/ndarray.html
#serialize-from-to-distributed-filesystems).