From 7afe6ba84e73205265150c1b91aa332a4e9aab01 Mon Sep 17 00:00:00 2001
From: Lianmin Zheng <mercy_zheng@sjtu.edu.cn>
Date: Thu, 23 Aug 2018 13:43:23 -0700
Subject: [PATCH] fix CO CI problem (#1641)
---
tutorials/autotvm/tune_conv2d_cuda.py | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/tutorials/autotvm/tune_conv2d_cuda.py b/tutorials/autotvm/tune_conv2d_cuda.py
index 375d1a9b7..3ff26a050 100644
--- a/tutorials/autotvm/tune_conv2d_cuda.py
+++ b/tutorials/autotvm/tune_conv2d_cuda.py
@@ -64,7 +64,7 @@ from tvm import autotvm
#
@autotvm.template
-def conv2d_no_batching(N, H, W, CI, CO, KH, KW, stride, padding):
+def conv2d_no_batching(N, H, W, CO, CI, KH, KW, stride, padding):
assert N == 1, "Only consider batch_size = 1 in this template"
data = tvm.placeholder((N, CI, H, W), name='data')
@@ -206,8 +206,8 @@ func(a_tvm, w_tvm, c_tvm)
np.testing.assert_allclose(c_np, c_tvm.asnumpy(), rtol=1e-2)
-# Evaluate running time. Here we choose a large repeat number (200) to reduce the noise
+# Evaluate running time. Here we choose a large repeat number (400) to reduce the noise
# and the overhead of kernel launch. You can also use nvprof to validate the result.
-evaluator = func.time_evaluator(func.entry_name, ctx, number=200)
+evaluator = func.time_evaluator(func.entry_name, ctx, number=400)
print('Time cost of this operator: %f' % evaluator(a_tvm, w_tvm, c_tvm).mean)
--
GitLab