From d6162af210d9d5648d33bf0fda40f773ac200df5 Mon Sep 17 00:00:00 2001 From: AlexeyAB <alexeyab84@gmail.com> Date: Wed, 08 Aug 2018 23:31:36 +0000 Subject: [PATCH] Optimized on CPU: gemm_bin, im2col, activation, transpose --- src/convolutional_layer.h | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/src/convolutional_layer.h b/src/convolutional_layer.h index 970aa10..b804afb 100644 --- a/src/convolutional_layer.h +++ b/src/convolutional_layer.h @@ -20,7 +20,8 @@ void add_bias_gpu(float *output, float *biases, int batch, int n, int size); void backward_bias_gpu(float *bias_updates, float *delta, int batch, int n, int size); #ifdef CUDNN -void cudnn_convolutional_setup(layer *l); +void cudnn_convolutional_setup(layer *l, int cudnn_preference); +void cuda_convert_f32_to_f16(float* input_f32, size_t size, float *output_f16); #endif #endif @@ -34,6 +35,8 @@ void swap_binary(convolutional_layer *l); void binarize_weights2(float *weights, int n, int size, char *binary, float *scales); +void binary_align_weights(convolutional_layer *l, size_t ldb_align); + void backward_convolutional_layer(convolutional_layer layer, network_state state); void add_bias(float *output, float *biases, int batch, int n, int size); -- Gitblit v1.10.0