You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit bfa7463
Browse filesBrowse the repository at this point in the historyBrowse files
Copy file name to clipboardExpand all lines: tests/py/dynamo/conversion/test_cumsum_aten.py
+28-1Lines changed: 28 additions & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -1,3 +1,5 @@
1
+
importunittest
2
+
1
3
importtorch
2
4
importtorch.nnasnn
3
5
importtorch_tensorrt
@@ -75,7 +77,6 @@ def forward(self, x):
75
77
((2, 3), (2, 4), (2, 5), 0),
76
78
((2, 3), (3, 4), (4, 5), -1),
77
79
((1, 2, 2), (2, 2, 3), (3, 3, 3), 0),
78
-
((1, 2, 2), (2, 2, 3), (3, 2, 3), -2),
79
80
((1, 2, 2, 3), (2, 2, 3, 4), (3, 3, 4, 5), -3),
80
81
((1, 2, 2, 3), (2, 2, 3, 4), (3, 3, 4, 5), -2),
81
82
]
@@ -98,6 +99,32 @@ def forward(self, x):
98
99
immutable_weights=False,
99
100
)
100
101
102
+
@unittest.skipIf(
103
+
torch_tensorrt.ENABLED_FEATURES.tensorrt_rtx,
104
+
"TRT RTX (CPU-only AoT build) cannot run ConstantBuilder, which TRT requires for loops whose trip count is a compile-time constant; occurs here because dim 1 has the same size across all profiles",
105
+
)
106
+
deftest_cumsum_dynamic_shape_static_dim(self):
107
+
# min=(1,2,2), opt=(2,2,3), max=(3,2,3), dim=-2=1
108
+
# dim 1 is size 2 in all profiles, so TRT uses a constant trip limit
0 commit comments