Tensorflow-style analyzer.
- Source:
Methods
(inner) analyzeConvTF(input, filter, strides, padding, options) → {Object}
- Source:
Calculate statistics for convolution layer. Check link for how to calculate output size.
Parameters:
| Name | Type | Description |
|---|---|---|
input |
Array | 2 or 3 dimension array ([height, width] or [height, width, channel]) |
filter |
Array | 3 dimension array ([height, width, out-channel]) |
strides |
Array | 2 dimension array ([height, width]) |
padding |
String | The padding algorithm. Either "VALID" or "SAME". |
options |
Object |
Returns:
statistics of the convolution layer
- Type
- Object
(inner) analyzePoolTF(input, filter, strides, padding, options) → {Object}
- Source:
Calculate statistics for pooling layer
Parameters:
| Name | Type | Description |
|---|---|---|
input |
Array | 2 or 3 dimension array ([height, width] or [height, width, channel]) |
filter |
Array | 2 dimension array ([height, width]) |
strides |
Array | 2 dimension array ([height, width]) |
padding |
String | The padding algorithm. Either "VALID" or "SAME". |
options |
Object |
Returns:
statistics of the pooling layer
- Type
- Object
(inner) getPaddings(input, filter, strides, options) → {Object}
- Source:
Calculate paddings for "SAME". Reference at https://www.tensorflow.org/api_guides/python/nn
Parameters:
| Name | Type | Description |
|---|---|---|
input |
Array | 2 or 3 dimension array ([height, width] or [height, width, channel]) |
filter |
Array | 2 dimension array ([height, width]) |
strides |
Array | 2 dimension array ([height, width]) |
options |
Object |
Returns:
statistics of the convolution layer
- Type
- Object