| cudnn_activation_backward | Calculates the gradient of the neural activation function
|
| cudnn_activation_forward | Applies a pointwise neural activation function to the input signal
|
| cudnn_convolution_backward_bias | Calculates the gradient of the convolution with respect to the bias, which is defined as the sum of all intensities per
feature map across all images of the input tensor.
|
| cudnn_convolution_backward_data | Calculates the gradient of the convolution with respect to the data using the CUDNN library
|
| cudnn_convolution_backward_filter | Calculates the gradient of the convolution with respect to the filter coefficients, using the CUDNN library.
|
| cudnn_convolution_forward | Calculates a forward convolution or cross-correlation using the CUDNN library.
|
| cudnn_hwcn_tensor | Converts a 4D tensor in nchw format to hwcn format (h=height, w=width, c=channels, n=images)
|
| cudnn_nchw_tensor | Converts a 4D tensor in hwcn format to nchw format (h=height, w=width, c=channels, n=images)
|
| cudnn_pooling_backward | Calculates the gradient of the pooling operation, using the CUDNN library
|
| cudnn_pooling_forward | Performs pooling of the input values (i.e., the maximum or average of a local neighborhood of values), resulting
in a tensor of smaller dimensions. The dimensions of this tensor can be retrieved using
|
| cudnn_pooling_getforwardoutputdim | Provides the output tensor dimensions after the pooling has been applied, using the following formula
which is applied to the spatial dimensions of the tensor.
|
| cudnn_softmax_backward | This function calculates the gradient of the softmax function, using the CUDNN library
|
| cudnn_softmax_forward | This function calculates the softmax function, using the CUDNN library
|
| cudnn_version | Returns the version of the currently installed cuDNN library.
|
| qdnn_activation_descriptor | The base class for all objects in Quasar |
| qdnn_conv_descriptor | The base class for all objects in Quasar |
| qdnn_pooling_descriptor | The base class for all objects in Quasar |