From ead150a2f00136f6c72dee81f17b7851fa4286d9 Mon Sep 17 00:00:00 2001
From: Max Weinzierl <supermaxman1@gmail.com>
Date: Thu, 18 Oct 2018 11:11:12 -0500
Subject: [PATCH] Fix for RuntimeError on torch tensor indexing

This resolves #2
Fix for RuntimeError: tensors used as indices must be long or byte tensors
---
 pytorch_neat/recurrent_net.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pytorch_neat/recurrent_net.py b/pytorch_neat/recurrent_net.py
index 1bf811b..413971e 100644
--- a/pytorch_neat/recurrent_net.py
+++ b/pytorch_neat/recurrent_net.py
@@ -33,7 +33,7 @@ def dense_from_coo(shape, conns, dtype=torch.float64):
     idxs, weights = conns
     if len(idxs) == 0:
         return mat
-    rows, cols = np.array(idxs).transpose()
+    rows, cols = np.array(idxs, dtype=np.int64).transpose()
     mat[torch.tensor(rows), torch.tensor(cols)] = torch.tensor(
         weights, dtype=dtype)
     return mat