remove @torch.jit.script annotations (move jit compatibility to the test suite)

Author: rusty1s

test/composite/test_logsumexp.py

@@ -18,3 +18,6 @@ def test_logsumexp():
         assert out.tolist() == torch.logsumexp(src, dim=0).tolist()
 
     outputs.backward(torch.randn_like(outputs))
+
+    jit = torch.jit.script(scatter_logsumexp)
+    assert jit(inputs, index).tolist() == outputs.tolist()

test/composite/test_softmax.py

@@ -22,6 +22,9 @@ def test_softmax():
 
     out.backward(torch.randn_like(out))
 
+    jit = torch.jit.script(scatter_softmax)
+    assert jit(src, index).tolist() == out.tolist()
+
 
 def test_log_softmax():
     src = torch.tensor([0.2, 0, 0.2, -2.1, 3.2, 7, -1, float('-inf')])
@@ -42,3 +45,6 @@ def test_log_softmax():
     assert torch.allclose(out, expected)
 
     out.backward(torch.randn_like(out))
+
+    jit = torch.jit.script(scatter_log_softmax)
+    assert jit(src, index).tolist() == out.tolist()

test/composite/test_std.py

@@ -13,3 +13,6 @@ def test_std():
     assert torch.allclose(out, expected)
 
     out.backward(torch.randn_like(out))
+
+    jit = torch.jit.script(scatter_std)
+    assert jit(src, index, dim=-1, unbiased=True).tolist() == out.tolist()

test/test_scatter.py

@@ -99,12 +99,18 @@ def test_forward(test, reduce, dtype, device):
     dim = test['dim']
     expected = tensor(test[reduce], dtype, device)
 
-    out = getattr(torch_scatter, 'scatter_' + reduce)(src, index, dim)
-    if isinstance(out, tuple):
-        out, arg_out = out
+    fn = getattr(torch_scatter, 'scatter_' + reduce)
+    jit = torch.jit.script(fn)
+    out1 = fn(src, index, dim)
+    out2 = jit(src, index, dim)
+    if isinstance(out1, tuple):
+        out1, arg_out1 = out1
+        out2, arg_out2 = out2
         arg_expected = tensor(test['arg_' + reduce], torch.long, device)
-        assert torch.all(arg_out == arg_expected)
-    assert torch.all(out == expected)
+        assert torch.all(arg_out1 == arg_expected)
+        assert arg_out1.tolist() == arg_out1.tolist()
+    assert torch.all(out1 == expected)
+    assert out1.tolist() == out2.tolist()
 
 
 @pytest.mark.parametrize('test,reduce,device',

test/test_segment.py

@@ -91,19 +91,31 @@ def test_forward(test, reduce, dtype, device):
     indptr = tensor(test['indptr'], torch.long, device)
     expected = tensor(test[reduce], dtype, device)
 
-    out = getattr(torch_scatter, 'segment_' + reduce + '_csr')(src, indptr)
-    if isinstance(out, tuple):
-        out, arg_out = out
+    fn = getattr(torch_scatter, 'segment_' + reduce + '_csr')
+    jit = torch.jit.script(fn)
+    out1 = fn(src, indptr)
+    out2 = jit(src, indptr)
+    if isinstance(out1, tuple):
+        out1, arg_out1 = out1
+        out2, arg_out2 = out2
         arg_expected = tensor(test['arg_' + reduce], torch.long, device)
-        assert torch.all(arg_out == arg_expected)
-    assert torch.all(out == expected)
-
-    out = getattr(torch_scatter, 'segment_' + reduce + '_coo')(src, index)
-    if isinstance(out, tuple):
-        out, arg_out = out
+        assert torch.all(arg_out1 == arg_expected)
+        assert arg_out1.tolist() == arg_out2.tolist()
+    assert torch.all(out1 == expected)
+    assert out1.tolist() == out2.tolist()
+
+    fn = getattr(torch_scatter, 'segment_' + reduce + '_coo')
+    jit = torch.jit.script(fn)
+    out1 = fn(src, index)
+    out2 = jit(src, index)
+    if isinstance(out1, tuple):
+        out1, arg_out1 = out1
+        out2, arg_out2 = out2
         arg_expected = tensor(test['arg_' + reduce], torch.long, device)
-        assert torch.all(arg_out == arg_expected)
-    assert torch.all(out == expected)
+        assert torch.all(arg_out1 == arg_expected)
+        assert arg_out1.tolist() == arg_out2.tolist()
+    assert torch.all(out1 == expected)
+    assert out1.tolist() == out2.tolist()
 
 
 @pytest.mark.parametrize('test,reduce,device',

torch_scatter/composite/logsumexp.py

@@ -6,7 +6,6 @@
 from torch_scatter.utils import broadcast
 
 
-@torch.jit.script
 def scatter_logsumexp(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                       out: Optional[torch.Tensor] = None,
                       dim_size: Optional[int] = None,

torch_scatter/composite/softmax.py

@@ -4,7 +4,6 @@
 from torch_scatter.utils import broadcast
 
 
-@torch.jit.script
 def scatter_softmax(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                     eps: float = 1e-12) -> torch.Tensor:
     if not torch.is_floating_point(src):
@@ -25,7 +24,6 @@ def scatter_softmax(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
     return recentered_scores_exp.div(normalizing_constants)
 
 
-@torch.jit.script
 def scatter_log_softmax(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                         eps: float = 1e-12) -> torch.Tensor:
     if not torch.is_floating_point(src):

torch_scatter/composite/std.py

@@ -5,7 +5,6 @@
 from torch_scatter.utils import broadcast
 
 
-@torch.jit.script
 def scatter_std(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                 out: Optional[torch.Tensor] = None,
                 dim_size: Optional[int] = None,

torch_scatter/scatter.py

@@ -5,7 +5,6 @@
 from .utils import broadcast
 
 
-@torch.jit.script
 def scatter_sum(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                 out: Optional[torch.Tensor] = None,
                 dim_size: Optional[int] = None) -> torch.Tensor:
@@ -24,21 +23,18 @@ def scatter_sum(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
         return out.scatter_add_(dim, index, src)
 
 
-@torch.jit.script
 def scatter_add(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                 out: Optional[torch.Tensor] = None,
                 dim_size: Optional[int] = None) -> torch.Tensor:
     return scatter_sum(src, index, dim, out, dim_size)
 
 
-@torch.jit.script
 def scatter_mul(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                 out: Optional[torch.Tensor] = None,
                 dim_size: Optional[int] = None) -> torch.Tensor:
     return torch.ops.torch_scatter.scatter_mul(src, index, dim, out, dim_size)
 
 
-@torch.jit.script
 def scatter_mean(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
                  out: Optional[torch.Tensor] = None,
                  dim_size: Optional[int] = None) -> torch.Tensor:
@@ -63,15 +59,13 @@ def scatter_mean(src: torch.Tensor, index: torch.Tensor, dim: int = -1,
     return out
 
 
-@torch.jit.script
 def scatter_min(
         src: torch.Tensor, index: torch.Tensor, dim: int = -1,
         out: Optional[torch.Tensor] = None,
         dim_size: Optional[int] = None) -> Tuple[torch.Tensor, torch.Tensor]:
     return torch.ops.torch_scatter.scatter_min(src, index, dim, out, dim_size)
 
 
-@torch.jit.script
 def scatter_max(
         src: torch.Tensor, index: torch.Tensor, dim: int = -1,
         out: Optional[torch.Tensor] = None,

torch_scatter/segment_coo.py

@@ -3,40 +3,35 @@
 import torch
 
 
-@torch.jit.script
 def segment_sum_coo(src: torch.Tensor, index: torch.Tensor,
                     out: Optional[torch.Tensor] = None,
                     dim_size: Optional[int] = None) -> torch.Tensor:
     return torch.ops.torch_scatter.segment_sum_coo(src, index, out, dim_size)
 
 
-@torch.jit.script
 def segment_add_coo(src: torch.Tensor, index: torch.Tensor,
                     out: Optional[torch.Tensor] = None,
                     dim_size: Optional[int] = None) -> torch.Tensor:
     return torch.ops.torch_scatter.segment_sum_coo(src, index, out, dim_size)
 
 
-@torch.jit.script
 def segment_mean_coo(src: torch.Tensor, index: torch.Tensor,
                      out: Optional[torch.Tensor] = None,
                      dim_size: Optional[int] = None) -> torch.Tensor:
     return torch.ops.torch_scatter.segment_mean_coo(src, index, out, dim_size)
 
 
-@torch.jit.script
-def segment_min_coo(src: torch.Tensor, index: torch.Tensor,
-                    out: Optional[torch.Tensor] = None,
-                    dim_size: Optional[int] = None
-                    ) -> Tuple[torch.Tensor, torch.Tensor]:
+def segment_min_coo(
+        src: torch.Tensor, index: torch.Tensor,
+        out: Optional[torch.Tensor] = None,
+        dim_size: Optional[int] = None) -> Tuple[torch.Tensor, torch.Tensor]:
     return torch.ops.torch_scatter.segment_min_coo(src, index, out, dim_size)
 
 
-@torch.jit.script
-def segment_max_coo(src: torch.Tensor, index: torch.Tensor,
-                    out: Optional[torch.Tensor] = None,
-                    dim_size: Optional[int] = None
-                    ) -> Tuple[torch.Tensor, torch.Tensor]:
+def segment_max_coo(
+        src: torch.Tensor, index: torch.Tensor,
+        out: Optional[torch.Tensor] = None,
+        dim_size: Optional[int] = None) -> Tuple[torch.Tensor, torch.Tensor]:
     return torch.ops.torch_scatter.segment_max_coo(src, index, out, dim_size)
 
 
@@ -137,7 +132,6 @@ def segment_coo(src: torch.Tensor, index: torch.Tensor,
         raise ValueError
 
 
-@torch.jit.script
 def gather_coo(src: torch.Tensor, index: torch.Tensor,
                out: Optional[torch.Tensor] = None) -> torch.Tensor:
     return torch.ops.torch_scatter.gather_coo(src, index, out)