move convert dtype to functional_tensor since only works on tensors

nairbv · nairbv · commit f610463f1aa8 · 2020-07-17T14:41:49.000-04:00
diff --git a/test/test_functional_tensor.py b/test/test_functional_tensor.py
@@ -309,7 +309,7 @@ def test_adjust_gamma(self):
         for dt in [torch.float64, torch.float32, None]:
 
             if dt is not None:
-                tensor = F.convert_image_dtype(tensor, dt)
+                tensor = F_t.convert_image_dtype(tensor, dt)
 
             gammas = [0.8, 1.0, 1.2]
             gains = [0.7, 1.0, 1.3]
@@ -323,7 +323,7 @@ def test_adjust_gamma(self):
 
                 rbg_tensor = adjusted_tensor
                 if adjusted_tensor.dtype != torch.uint8:
-                    rbg_tensor = F.convert_image_dtype(adjusted_tensor, torch.uint8)
+                    rbg_tensor = F_t.convert_image_dtype(adjusted_tensor, torch.uint8)
 
                 self.compareTensorToPIL(rbg_tensor, adjusted_pil)
 
diff --git a/test/test_transforms.py b/test/test_transforms.py
@@ -2,6 +2,7 @@
 import torch
 import torchvision.transforms as transforms
 import torchvision.transforms.functional as F
+import torchvision.transforms.functional_tensor as F_t
 from torch._utils_internal import get_file_path_2
 from numpy.testing import assert_array_almost_equal
 import unittest
@@ -528,7 +529,7 @@ def test_to_tensor(self):
 
     def test_max_value(self):
         for dtype in int_dtypes():
-            self.assertEqual(F._max_value(dtype), torch.iinfo(dtype).max)
+            self.assertEqual(F_t._max_value(dtype), torch.iinfo(dtype).max)
 
     def test_convert_image_dtype_float_to_float(self):
         for input_dtype, output_dtypes in cycle_over(float_dtypes()):
diff --git a/torchvision/transforms/functional.py b/torchvision/transforms/functional.py
@@ -124,88 +124,6 @@ def pil_to_tensor(pic):
     return img
 
 
-# torch.iinfo isn't scriptable so using this helper function
-# https://github.com/pytorch/pytorch/issues/41492
-def _max_value(dtype: int) -> int:
-    a = torch.tensor(2, dtype=dtype)
-    signed = 1 if torch.tensor(0, dtype=dtype).is_signed() else 0
-    bits = 1
-    max_value = torch.tensor(-signed, dtype=torch.long)
-    while(True):
-        next_value = a.pow(bits - signed).sub(1)
-        if next_value > max_value:
-            max_value = next_value
-            bits *= 2
-        else:
-            return max_value.item()
-    return max_value.item()
-
-
-def convert_image_dtype(image: torch.Tensor, dtype: int = torch.float) -> torch.Tensor:
-    """Convert a tensor image to the given ``dtype`` and scale the values accordingly
-
-    Args:
-        image (torch.Tensor): Image to be converted
-        dtype (torch.dtype): Desired data type of the output
-
-    Returns:
-        (torch.Tensor): Converted image
-
-    .. note::
-
-        When converting from a smaller to a larger integer ``dtype`` the maximum values are **not** mapped exactly.
-        If converted back and forth, this mismatch has no effect.
-
-    Raises:
-        RuntimeError: When trying to cast :class:`torch.float32` to :class:`torch.int32` or :class:`torch.int64` as
-            well as for trying to cast :class:`torch.float64` to :class:`torch.int64`. These conversions might lead to
-            overflow errors since the floating point ``dtype`` cannot store consecutive integers over the whole range
-            of the integer ``dtype``.
-    """
-    if image.dtype == dtype:
-        return image
-
-    if torch.empty(0, dtype=image.dtype).is_floating_point():
-        # float to float
-        if torch.tensor(0, dtype=dtype).is_floating_point():
-            return image.to(dtype)
-
-        # float to int
-        if (image.dtype == torch.float32 and dtype in (torch.int32, torch.int64)) or (
-            image.dtype == torch.float64 and dtype == torch.int64
-        ):
-            msg = f"The cast from {image.dtype} to {dtype} cannot be performed safely."
-            raise RuntimeError(msg)
-
-        # https://github.com/pytorch/vision/pull/2078#issuecomment-612045321
-        # For data in the range 0-1, (float * 255).to(uint) is only 255
-        # when float is exactly 1.0.
-        # `max + 1 - epsilon` provides more evenly distributed mapping of
-        # ranges of floats to ints.
-        eps = 1e-3
-        max_val = _max_value(dtype)
-        result = image.mul(max_val + 1.0 - eps)
-        return result.to(dtype)
-    else:
-        input_max = _max_value(image.dtype)
-        output_max = _max_value(dtype)
-
-        # int to float
-        if torch.tensor(0, dtype=dtype).is_floating_point():
-            image = image.to(dtype)
-            return image / input_max
-
-        # int to int
-        if input_max > output_max:
-            factor = (input_max + 1) // (output_max + 1)
-            image = image // factor
-            return image.to(dtype)
-        else:
-            factor = (output_max + 1) // (input_max + 1)
-            image = image.to(dtype)
-            return image * factor
-
-
 def to_pil_image(pic, mode=None):
     """Convert a tensor or an ndarray to PIL Image.
 
diff --git a/torchvision/transforms/functional_tensor.py b/torchvision/transforms/functional_tensor.py
@@ -6,8 +6,6 @@
 from torch.nn.functional import affine_grid, grid_sample
 from torch.jit.annotations import List, BroadcastingList2
 
-import torchvision.transforms.functional as F
-
 
 def _is_tensor_a_torch_image(x: Tensor) -> bool:
     return x.ndim >= 2
@@ -20,6 +18,88 @@ def _get_image_size(img: Tensor) -> List[int]:
     raise TypeError("Unexpected type {}".format(type(img)))
 
 
+# torch.iinfo isn't scriptable so using this helper function
+# https://github.com/pytorch/pytorch/issues/41492
+def _max_value(dtype: int) -> int:
+    a = torch.tensor(2, dtype=dtype)
+    signed = 1 if torch.tensor(0, dtype=dtype).is_signed() else 0
+    bits = 1
+    max_value = torch.tensor(-signed, dtype=torch.long)
+    while(True):
+        next_value = a.pow(bits - signed).sub(1)
+        if next_value > max_value:
+            max_value = next_value
+            bits *= 2
+        else:
+            return max_value.item()
+    return max_value.item()
+
+
+def convert_image_dtype(image: torch.Tensor, dtype: int = torch.float) -> torch.Tensor:
+    """Convert a tensor image to the given ``dtype`` and scale the values accordingly
+
+    Args:
+        image (torch.Tensor): Image to be converted
+        dtype (torch.dtype): Desired data type of the output
+
+    Returns:
+        (torch.Tensor): Converted image
+
+    .. note::
+
+        When converting from a smaller to a larger integer ``dtype`` the maximum values are **not** mapped exactly.
+        If converted back and forth, this mismatch has no effect.
+
+    Raises:
+        RuntimeError: When trying to cast :class:`torch.float32` to :class:`torch.int32` or :class:`torch.int64` as
+            well as for trying to cast :class:`torch.float64` to :class:`torch.int64`. These conversions might lead to
+            overflow errors since the floating point ``dtype`` cannot store consecutive integers over the whole range
+            of the integer ``dtype``.
+    """
+    if image.dtype == dtype:
+        return image
+
+    if torch.empty(0, dtype=image.dtype).is_floating_point():
+        # float to float
+        if torch.tensor(0, dtype=dtype).is_floating_point():
+            return image.to(dtype)
+
+        # float to int
+        if (image.dtype == torch.float32 and dtype in (torch.int32, torch.int64)) or (
+            image.dtype == torch.float64 and dtype == torch.int64
+        ):
+            msg = f"The cast from {image.dtype} to {dtype} cannot be performed safely."
+            raise RuntimeError(msg)
+
+        # https://github.com/pytorch/vision/pull/2078#issuecomment-612045321
+        # For data in the range 0-1, (float * 255).to(uint) is only 255
+        # when float is exactly 1.0.
+        # `max + 1 - epsilon` provides more evenly distributed mapping of
+        # ranges of floats to ints.
+        eps = 1e-3
+        max_val = _max_value(dtype)
+        result = image.mul(max_val + 1.0 - eps)
+        return result.to(dtype)
+    else:
+        input_max = _max_value(image.dtype)
+        output_max = _max_value(dtype)
+
+        # int to float
+        if torch.tensor(0, dtype=dtype).is_floating_point():
+            image = image.to(dtype)
+            return image / input_max
+
+        # int to int
+        if input_max > output_max:
+            factor = (input_max + 1) // (output_max + 1)
+            image = image // factor
+            return image.to(dtype)
+        else:
+            factor = (output_max + 1) // (input_max + 1)
+            image = image.to(dtype)
+            return image * factor
+
+
 def vflip(img: Tensor) -> Tensor:
     """Vertically flip the given the Image Tensor.
 
@@ -230,11 +310,11 @@ def adjust_gamma(img: Tensor, gamma: float, gain: float = 1) -> Tensor:
     result = img
     dtype = img.dtype
     if not torch.is_floating_point(img):
-        result = F.convert_image_dtype(result, torch.get_default_dtype())
+        result = convert_image_dtype(result, torch.float32)
 
     result = (gain * result ** gamma).clamp(0, 1)
 
-    result = F.convert_image_dtype(result, dtype)
+    result = convert_image_dtype(result, dtype)
     result = result.to(dtype)
     return result
 
diff --git a/torchvision/transforms/transforms.py b/torchvision/transforms/transforms.py
@@ -16,7 +16,7 @@
     accimage = None
 
 from . import functional as F
-
+from . import functional_tensor as F_t
 
 __all__ = ["Compose", "ToTensor", "PILToTensor", "ConvertImageDtype", "ToPILImage", "Normalize", "Resize", "Scale",
            "CenterCrop", "Pad", "Lambda", "RandomApply", "RandomChoice", "RandomOrder", "RandomCrop",
@@ -131,7 +131,7 @@ def __init__(self, dtype: torch.dtype) -> None:
         self.dtype = dtype
 
     def __call__(self, image: torch.Tensor) -> torch.Tensor:
-        return F.convert_image_dtype(image, self.dtype)
+        return F_t.convert_image_dtype(image, self.dtype)
 
 
 class ToPILImage(object):
