From 9bae70b22549b68f5cdeece8b6c3b3de00c22714 Mon Sep 17 00:00:00 2001
From: AlexeyAB <alexeyab84@gmail.com>
Date: Mon, 16 Apr 2018 23:51:11 +0000
Subject: [PATCH] Accelerated by another 5% using FP16/32 Batch-norm for Tensor Cores.

---
 src/convolutional_layer.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/src/convolutional_layer.h b/src/convolutional_layer.h
index da98dce..6d1e517 100644
--- a/src/convolutional_layer.h
+++ b/src/convolutional_layer.h
@@ -21,6 +21,7 @@
 void backward_bias_gpu(float *bias_updates, float *delta, int batch, int n, int size);
 #ifdef CUDNN
 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
 

--
Gitblit v1.10.0