From cad4d1618fee74471d335314cb77070fee951a42 Mon Sep 17 00:00:00 2001
From: AlexeyAB <alexeyab84@gmail.com>
Date: Sun, 25 Feb 2018 13:29:44 +0000
Subject: [PATCH] Added support for Tensor Cores CC >= 7.0 (V100). For FP16/32 (mixed precision) define CUDNN_HALF should be used.
---
src/rnn_layer.c | 6 ++++++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/src/rnn_layer.c b/src/rnn_layer.c
index b713899..83fda13 100644
--- a/src/rnn_layer.c
+++ b/src/rnn_layer.c
@@ -58,7 +58,13 @@
l.output = l.output_layer->output;
l.delta = l.output_layer->delta;
+ l.forward = forward_rnn_layer;
+ l.backward = backward_rnn_layer;
+ l.update = update_rnn_layer;
#ifdef GPU
+ l.forward_gpu = forward_rnn_layer_gpu;
+ l.backward_gpu = backward_rnn_layer_gpu;
+ l.update_gpu = update_rnn_layer_gpu;
l.state_gpu = cuda_make_array(l.state, batch*hidden*(steps+1));
l.output_gpu = l.output_layer->output_gpu;
l.delta_gpu = l.output_layer->delta_gpu;
--
Gitblit v1.10.0