Add input/outputShape information to all KDocs of layers (#211) #460
Conversation
� Conflicts: � api/src/main/kotlin/org/jetbrains/kotlinx/dl/api/inference/keras/ModelSaver.kt
- added emphasis - removed redundant note comment
ermolenkodev
requested review from
ermolenkodev
and removed request for
juliabeliaeva
There was a problem hiding this comment.
Sorry for such a late review.
There are a few minor issues with the docs.
You make the following comment in many places:
Use the keyword argument 'input_shape'
(tuple of integers, does not include the samples axis)
when using this layer as the first layer in a model.
But currently, we only support Input to be the first layer of the network. Moreover, there is no argument named input_shape in the constructor anywhere in the code.
tensorflow/src/main/kotlin/org/jetbrains/kotlinx/dl/api/core/layer/core/Dense.kt
Outdated
Show resolved
Hide resolved
| @@ -10,6 +10,11 @@ import org.tensorflow.Shape | |||
| /** | |||
| * Zero-padding layer for 1D input (e.g. audio). | |||
| * This layer can add zeros in the rows of the audio tensor | |||
| * | |||
| * __Input shape:__ 3D tensor with shape `(batch_size, axis_to_pad, features)`. | |||
There was a problem hiding this comment.
axis_to_pad is a somewhat misleading name because an axis usually refers to the dimension's index, but here it refers to the size of the dimension. Just dim would be ok.
| @@ -10,6 +10,11 @@ import org.tensorflow.Shape | |||
| /** | |||
| * Zero-padding layer for 3D input (e.g. video). | |||
| * This layer can add zeros in the rows, cols and depth of a video tensor. | |||
| * | |||
| * __Input shape:__ 5D tensor with shape `(batch_size, first_axis_to_pad, second_axis_to_pad, third_axis_to_pad, depth)`. | |||
There was a problem hiding this comment.
Again, maybe it would be better to replace axis with dim
I see, thank you for the feedback! Also my apologies for the late reply, I have been quite busy. |
Changed N-D to 2D since only 2D input is supported for now.
No description provided.