From a284a7da8d1facbf984a22302665a2a50295a687 Mon Sep 17 00:00:00 2001
From: AlexeyAB <alexeyab84@gmail.com>
Date: Wed, 08 Aug 2018 16:08:58 +0000
Subject: [PATCH] Try to use avx_hs() - slow and requires alignment 4096 bits < (l.size*l.size*l.c) May be faster only from 8192 bits and more.

---
 src/convolutional_layer.h |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/src/convolutional_layer.h b/src/convolutional_layer.h
index b7953ee..dd79c48 100644
--- a/src/convolutional_layer.h
+++ b/src/convolutional_layer.h
@@ -20,11 +20,12 @@
 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
 
-convolutional_layer make_convolutional_layer(int batch, int h, int w, int c, int n, int size, int stride, int pad, ACTIVATION activation, int batch_normalization, int binary, int xnor);
+convolutional_layer make_convolutional_layer(int batch, int h, int w, int c, int n, int size, int stride, int padding, ACTIVATION activation, int batch_normalize, int binary, int xnor, int adam);
 void denormalize_convolutional_layer(convolutional_layer l);
 void resize_convolutional_layer(convolutional_layer *layer, int w, int h);
 void forward_convolutional_layer(const convolutional_layer layer, network_state state);
@@ -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_transpose_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