From a296f11911284adb4ea0aaed8e9a16f925a27a03 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Thu, 25 Jan 2024 12:23:30 -0800
Subject: [PATCH 01/14] WIP. Tests failing.

---
 test/test_transforms_v2.py                    | 26 +++++++++++++++++
 torchvision/transforms/_functional_pil.py     |  7 +++++
 torchvision/transforms/_functional_tensor.py  | 13 +++++++++
 torchvision/transforms/functional.py          | 20 +++++++++++++
 torchvision/transforms/transforms.py          | 28 +++++++++++++++++++
 torchvision/transforms/v2/__init__.py         |  1 +
 torchvision/transforms/v2/_color.py           | 16 +++++++++++
 .../transforms/v2/functional/__init__.py      |  2 ++
 .../transforms/v2/functional/_color.py        | 22 +++++++++++++++
 9 files changed, 135 insertions(+)

diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
index 831a7e3b570..f7acb22eccb 100644
--- a/test/test_transforms_v2.py
+++ b/test/test_transforms_v2.py
@@ -4963,6 +4963,32 @@ def test_random_transform_correctness(self, num_input_channels):
         assert_equal(actual, expected, rtol=0, atol=1)
 
 
+class TestGrayscaleToRgb:
+    @pytest.mark.parametrize("dtype", [torch.uint8, torch.float32])
+    @pytest.mark.parametrize("device", cpu_and_cuda())
+    def test_kernel_image(self, dtype, device):
+        check_kernel(F.grayscale_to_rgb_image, make_image(dtype=dtype, device=device))
+
+    @pytest.mark.parametrize("make_input", [make_image_tensor, make_image_pil, make_image])
+    def test_functional(self, make_input):
+        check_functional(F.grayscale_to_rgb, make_input())
+
+    @pytest.mark.parametrize(
+        ("kernel", "input_type"),
+        [
+            (F.rgb_to_grayscale_image, torch.Tensor),
+            (F._rgb_to_grayscale_image_pil, PIL.Image.Image),
+            (F.rgb_to_grayscale_image, tv_tensors.Image),
+        ],
+    )
+    def test_functional_signature(self, kernel, input_type):
+        check_functional_kernel_signature_match(F.grayscale_to_rgb, kernel=kernel, input_type=input_type)
+
+    @pytest.mark.parametrize("make_input", [make_image_tensor, make_image_pil, make_image])
+    def test_transform(self, make_input):
+        check_transform(transforms.GrayscaleToRgb(), make_input(color_space="GRAY"))
+
+
 class TestRandomZoomOut:
     # Tests are light because this largely relies on the already tested `pad` kernels.
 
diff --git a/torchvision/transforms/_functional_pil.py b/torchvision/transforms/_functional_pil.py
index 277848224ac..f82684fafb9 100644
--- a/torchvision/transforms/_functional_pil.py
+++ b/torchvision/transforms/_functional_pil.py
@@ -348,6 +348,13 @@ def to_grayscale(img: Image.Image, num_output_channels: int) -> Image.Image:
     return img
 
 
+def grayscale_to_rgb(img: Image.Image) -> Image.Image:
+    if not _is_pil_image(img):
+        raise TypeError(f"img should be PIL Image. Got {type(img)}")
+
+    return img.convert("RGB")
+
+
 @torch.jit.unused
 def invert(img: Image.Image) -> Image.Image:
     if not _is_pil_image(img):
diff --git a/torchvision/transforms/_functional_tensor.py b/torchvision/transforms/_functional_tensor.py
index 88dc9ca21cc..9bccdfe5d9f 100644
--- a/torchvision/transforms/_functional_tensor.py
+++ b/torchvision/transforms/_functional_tensor.py
@@ -166,6 +166,19 @@ def rgb_to_grayscale(img: Tensor, num_output_channels: int = 1) -> Tensor:
     return l_img
 
 
+def grayscale_to_rgb(img: Tensor) -> Tensor:
+    if img.ndim < 3:
+        raise TypeError(f"Input image tensor should have at least 3 dimensions, but found {img.ndim}")
+    _assert_channels(img, [1, 3])
+
+    if img.shape[-3] != 1:
+        raise ValueError("Expected last dimension of input image tensor to be 1, got {}".format(img.shape[-3]))
+
+    s = [-1] * len(img.shape)
+    s[-3] = 3
+    return img.expand(s)
+
+
 def adjust_brightness(img: Tensor, brightness_factor: float) -> Tensor:
     if brightness_factor < 0:
         raise ValueError(f"brightness_factor ({brightness_factor}) is not non-negative.")
diff --git a/torchvision/transforms/functional.py b/torchvision/transforms/functional.py
index 698942c56af..034466d1316 100644
--- a/torchvision/transforms/functional.py
+++ b/torchvision/transforms/functional.py
@@ -1276,6 +1276,26 @@ def rgb_to_grayscale(img: Tensor, num_output_channels: int = 1) -> Tensor:
     return F_t.rgb_to_grayscale(img, num_output_channels)
 
 
+def grayscale_to_rgb(img: Tensor) -> Tensor:
+    """Converts grayscale image to 3-channel image by duplicating the value in all three channels.
+    If the image is torch Tensor, it is expected
+    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
+
+    Note:
+        Please, note that this method supports only grayscale images as input. For inputs in other color spaces,
+        please, consider using :meth:`~torchvision.transforms.functional.to_grayscale` with PIL Image.
+
+    Returns:
+        PIL Image or Tensor: RGB Image.
+    """
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(grayscale_to_rgb)
+    if not isinstance(img, torch.Tensor):
+        return F_pil.grayscale_to_rgb(img)
+    return F_t.grayscale_to_rgb(img)
+
+
+
 def erase(img: Tensor, i: int, j: int, h: int, w: int, v: Tensor, inplace: bool = False) -> Tensor:
     """Erase the input Tensor Image with given value.
     This transform does not support PIL Image.
diff --git a/torchvision/transforms/transforms.py b/torchvision/transforms/transforms.py
index 2a6e0ce12c0..8ec3e3466af 100644
--- a/torchvision/transforms/transforms.py
+++ b/torchvision/transforms/transforms.py
@@ -42,6 +42,7 @@
     "RandomRotation",
     "RandomAffine",
     "Grayscale",
+    "GrayscaleToRgb",
     "RandomGrayscale",
     "RandomPerspective",
     "RandomErasing",
@@ -1576,6 +1577,33 @@ def __repr__(self) -> str:
         return f"{self.__class__.__name__}(num_output_channels={self.num_output_channels})"
 
 
+class GrayscaleToRgb(torch.nn.Module):
+    """Converts grayscale image to 3 channel RGB image.
+    If the image is torch Tensor, it is expected
+    to have [..., 1, H, W] shape, where ... means an arbitrary number of leading dimensions
+
+    Returns:
+        PIL Image: RGB version of the input image.
+    """
+
+    def __init__(self):
+        super().__init__()
+        _log_api_usage_once(self)
+
+    def forward(self, img):
+        """
+        Args:
+            img (PIL Image or Tensor): Image to be converted to grayscale.
+
+        Returns:
+            PIL Image or Tensor: Grayscaled image.
+        """
+        return F.grayscale_to_rgb(img)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}()"
+
+
 class RandomGrayscale(torch.nn.Module):
     """Randomly convert image to grayscale with a probability of p (default 0.1).
     If the image is torch Tensor, it is expected
diff --git a/torchvision/transforms/v2/__init__.py b/torchvision/transforms/v2/__init__.py
index dbc0474d307..2cb3ea3eeae 100644
--- a/torchvision/transforms/v2/__init__.py
+++ b/torchvision/transforms/v2/__init__.py
@@ -9,6 +9,7 @@
 from ._color import (
     ColorJitter,
     Grayscale,
+    GrayscaleToRgb,
     RandomAdjustSharpness,
     RandomAutocontrast,
     RandomChannelPermutation,
diff --git a/torchvision/transforms/v2/_color.py b/torchvision/transforms/v2/_color.py
index d20953451ab..39605f46ed7 100644
--- a/torchvision/transforms/v2/_color.py
+++ b/torchvision/transforms/v2/_color.py
@@ -54,6 +54,22 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
         return self._call_kernel(F.rgb_to_grayscale, inpt, num_output_channels=params["num_input_channels"])
 
 
+class GrayscaleToRgb(Transform):
+    """Converts grayscale images to RGB images.
+
+    If the input is a :class:`torch.Tensor`, it is expected
+    to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions
+    """
+
+    _v1_transform_cls = _transforms.GrayscaleToRgb
+
+    def __init__(self):
+        super().__init__()
+
+    def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
+        return self._call_kernel(F.grayscale_to_rgb, inpt)
+
+
 class ColorJitter(Transform):
     """Randomly change the brightness, contrast, saturation and hue of an image or video.
 
diff --git a/torchvision/transforms/v2/functional/__init__.py b/torchvision/transforms/v2/functional/__init__.py
index 81d5c1b9baf..82f1d69cbba 100644
--- a/torchvision/transforms/v2/functional/__init__.py
+++ b/torchvision/transforms/v2/functional/__init__.py
@@ -63,6 +63,8 @@
     equalize,
     equalize_image,
     equalize_video,
+    grayscale_to_rgb,
+    grayscale_to_rgb_image,
     invert,
     invert_image,
     invert_video,
diff --git a/torchvision/transforms/v2/functional/_color.py b/torchvision/transforms/v2/functional/_color.py
index b0189fd95ef..36e0029a3c0 100644
--- a/torchvision/transforms/v2/functional/_color.py
+++ b/torchvision/transforms/v2/functional/_color.py
@@ -61,6 +61,28 @@ def _rgb_to_grayscale_image_pil(image: PIL.Image.Image, num_output_channels: int
     return _FP.to_grayscale(image, num_output_channels=num_output_channels)
 
 
+def grayscale_to_rgb(inpt: torch.Tensor) -> torch.Tensor:
+    """See :class:`~torchvision.transforms.v2.GrayscaleToRgb` for details."""
+    if torch.jit.is_scripting():
+        return grayscale_to_rgb_image(inpt)
+
+    _log_api_usage_once(grayscale_to_rgb)
+
+    kernel = _get_kernel(rgb_to_grayscale, type(inpt))
+    return kernel(inpt)
+
+
+@_register_kernel_internal(grayscale_to_rgb, torch.Tensor)
+@_register_kernel_internal(grayscale_to_rgb, tv_tensors.Image)
+def grayscale_to_rgb_image(image: torch.Tensor) -> torch.Tensor:
+    return _rgb_to_grayscale_image(image, num_output_channels=3, preserve_dtype=True)
+
+
+@_register_kernel_internal(grayscale_to_rgb, PIL.Image.Image)
+def grayscale_to_rgb_image(image: PIL.Image.Image) -> PIL.Image.Image:
+    return _FP.to_grayscale(image, num_output_channels=3)
+
+
 def _blend(image1: torch.Tensor, image2: torch.Tensor, ratio: float) -> torch.Tensor:
     ratio = float(ratio)
     fp = image1.is_floating_point()

From ff7a49c12bc788fa588feaf79dd4607cb159d5ec Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Thu, 1 Feb 2024 15:16:46 -0800
Subject: [PATCH 02/14] Fixed typo

---
 torchvision/transforms/v2/functional/_color.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/torchvision/transforms/v2/functional/_color.py b/torchvision/transforms/v2/functional/_color.py
index 3e6ac96640e..855c387e21e 100644
--- a/torchvision/transforms/v2/functional/_color.py
+++ b/torchvision/transforms/v2/functional/_color.py
@@ -83,7 +83,7 @@ def grayscale_to_rgb_image(image: torch.Tensor) -> torch.Tensor:
 
 
 @_register_kernel_internal(grayscale_to_rgb, PIL.Image.Image)
-def grayscale_to_rgb_image(image: PIL.Image.Image) -> PIL.Image.Image:
+def grayscale_to_rgb_image_pil(image: PIL.Image.Image) -> PIL.Image.Image:
     return _FP.to_grayscale(image, num_output_channels=3)
 
 

From 3f1e6114162a3b6828051655d52b9cc33e2b09ab Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Thu, 1 Feb 2024 15:25:43 -0800
Subject: [PATCH 03/14] Fixed typo

---
 test/test_transforms_v2.py | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
index ef5f32844ba..1334f0dbff2 100644
--- a/test/test_transforms_v2.py
+++ b/test/test_transforms_v2.py
@@ -4997,6 +4997,15 @@ def test_functional_signature(self, kernel, input_type):
     def test_transform(self, make_input):
         check_transform(transforms.GrayscaleToRgb(), make_input(color_space="GRAY"))
 
+    @pytest.mark.parametrize("fn", [F.grayscale_to_rgb, transform_cls_to_functional(transforms.GrayscaleToRgb)])
+    def test_image_correctness(self, fn):
+        image = make_image(dtype=torch.uint8, device="cpu", color_space="GRAY")
+
+        actual = fn(image, num_output_channels=num_output_channels)
+        expected = F.to_image(F.grayscale_to_rgb(F.to_pil_image(image)))
+
+        assert_equal(actual, expected, rtol=0, atol=1)
+
 
 class TestRandomZoomOut:
     # Tests are light because this largely relies on the already tested `pad` kernels.

From e0e056f74c14118b591be58936c70807d3b645b5 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Thu, 1 Feb 2024 15:45:11 -0800
Subject: [PATCH 04/14] Fixed bug.

---
 test/test_transforms_v2.py                     | 3 ++-
 torchvision/transforms/v2/functional/_color.py | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
index 1334f0dbff2..7f092c8115a 100644
--- a/test/test_transforms_v2.py
+++ b/test/test_transforms_v2.py
@@ -5001,9 +5001,10 @@ def test_transform(self, make_input):
     def test_image_correctness(self, fn):
         image = make_image(dtype=torch.uint8, device="cpu", color_space="GRAY")
 
-        actual = fn(image, num_output_channels=num_output_channels)
+        actual = fn(image)
         expected = F.to_image(F.grayscale_to_rgb(F.to_pil_image(image)))
 
+        print(f"ahmad: {expected.shape=} {actual.shape=}")
         assert_equal(actual, expected, rtol=0, atol=1)
 
 
diff --git a/torchvision/transforms/v2/functional/_color.py b/torchvision/transforms/v2/functional/_color.py
index 855c387e21e..f2e1d9ad06a 100644
--- a/torchvision/transforms/v2/functional/_color.py
+++ b/torchvision/transforms/v2/functional/_color.py
@@ -72,7 +72,7 @@ def grayscale_to_rgb(inpt: torch.Tensor) -> torch.Tensor:
 
     _log_api_usage_once(grayscale_to_rgb)
 
-    kernel = _get_kernel(rgb_to_grayscale, type(inpt))
+    kernel = _get_kernel(grayscale_to_rgb, type(inpt))
     return kernel(inpt)
 
 

From 3547f83c4ac6eaf62fd731713c53ac1a1f72f997 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Thu, 1 Feb 2024 15:45:43 -0800
Subject: [PATCH 05/14] Fixed bug.

---
 test/test_transforms_v2.py | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
index 7f092c8115a..fb8b5be2c48 100644
--- a/test/test_transforms_v2.py
+++ b/test/test_transforms_v2.py
@@ -5007,6 +5007,14 @@ def test_image_correctness(self, fn):
         print(f"ahmad: {expected.shape=} {actual.shape=}")
         assert_equal(actual, expected, rtol=0, atol=1)
 
+    def test_expanded_channels_are_not_views_into_the_same_underlying_tensor(self):
+        image = make_image(dtype=torch.uint8, device="cpu", color_space="GRAY")
+
+        output_image = F.grayscale_to_rgb(image)
+        assert_equal(output_image[0][0][0], output_image[1][0][0])
+        output_image[0][0][0] = output_image[0][0][0] + 1
+        assert output_image[0][0][0] != output_image[1][0][0]
+
 
 class TestRandomZoomOut:
     # Tests are light because this largely relies on the already tested `pad` kernels.

From 133a6ba1b82e3c5d0c313cf637dc20adb2e1472e Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Thu, 1 Feb 2024 15:48:58 -0800
Subject: [PATCH 06/14] Add comments.

---
 torchvision/transforms/v2/functional/_color.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/torchvision/transforms/v2/functional/_color.py b/torchvision/transforms/v2/functional/_color.py
index f2e1d9ad06a..fb1f9a4d174 100644
--- a/torchvision/transforms/v2/functional/_color.py
+++ b/torchvision/transforms/v2/functional/_color.py
@@ -79,11 +79,13 @@ def grayscale_to_rgb(inpt: torch.Tensor) -> torch.Tensor:
 @_register_kernel_internal(grayscale_to_rgb, torch.Tensor)
 @_register_kernel_internal(grayscale_to_rgb, tv_tensors.Image)
 def grayscale_to_rgb_image(image: torch.Tensor) -> torch.Tensor:
+    # rgb_to_grayscale can be used to add channels so we reuse that function.
     return _rgb_to_grayscale_image(image, num_output_channels=3, preserve_dtype=True)
 
 
 @_register_kernel_internal(grayscale_to_rgb, PIL.Image.Image)
 def grayscale_to_rgb_image_pil(image: PIL.Image.Image) -> PIL.Image.Image:
+    # to_grayscale can expand channels from 1 to 3 so we reuse that function.
     return _FP.to_grayscale(image, num_output_channels=3)
 
 

From 78db2f656dda508afbdc9a755c41ed9e5cb5cf86 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Thu, 1 Feb 2024 15:59:22 -0800
Subject: [PATCH 07/14] removed extra line

---
 torchvision/transforms/functional.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/torchvision/transforms/functional.py b/torchvision/transforms/functional.py
index 034466d1316..c83bf018adc 100644
--- a/torchvision/transforms/functional.py
+++ b/torchvision/transforms/functional.py
@@ -1295,7 +1295,6 @@ def grayscale_to_rgb(img: Tensor) -> Tensor:
     return F_t.grayscale_to_rgb(img)
 
 
-
 def erase(img: Tensor, i: int, j: int, h: int, w: int, v: Tensor, inplace: bool = False) -> Tensor:
     """Erase the input Tensor Image with given value.
     This transform does not support PIL Image.

From 990fc317864bd9da433a5fe22d18f86dc373ddc7 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Mon, 12 Feb 2024 11:23:47 -0800
Subject: [PATCH 08/14] Addressed comments and added another test.

---
 test/test_transforms_v2.py                     | 14 +++++++++++++-
 torchvision/transforms/v2/functional/_color.py |  6 ++++--
 2 files changed, 17 insertions(+), 3 deletions(-)

diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
index fb8b5be2c48..2af4dd9d739 100644
--- a/test/test_transforms_v2.py
+++ b/test/test_transforms_v2.py
@@ -5004,7 +5004,6 @@ def test_image_correctness(self, fn):
         actual = fn(image)
         expected = F.to_image(F.grayscale_to_rgb(F.to_pil_image(image)))
 
-        print(f"ahmad: {expected.shape=} {actual.shape=}")
         assert_equal(actual, expected, rtol=0, atol=1)
 
     def test_expanded_channels_are_not_views_into_the_same_underlying_tensor(self):
@@ -5015,6 +5014,19 @@ def test_expanded_channels_are_not_views_into_the_same_underlying_tensor(self):
         output_image[0][0][0] = output_image[0][0][0] + 1
         assert output_image[0][0][0] != output_image[1][0][0]
 
+    def test_rgb_image_is_unchanged(self):
+        image = make_image(dtype=torch.uint8, device="cpu", color_space="RGB")
+        assert_equal(image.shape[-3], 3)
+        image[0][0][0] = 0
+        image[1][0][0] = 100
+        image[2][0][0] = 200
+        output_image = F.grayscale_to_rgb(image)
+        assert output_image[0][0][0] == 0
+        assert output_image[1][0][0] == 100
+        assert output_image[2][0][0] == 200
+        print(image)
+        print(output_image)
+
 
 class TestRandomZoomOut:
     # Tests are light because this largely relies on the already tested `pad` kernels.
diff --git a/torchvision/transforms/v2/functional/_color.py b/torchvision/transforms/v2/functional/_color.py
index fb1f9a4d174..3025f876dff 100644
--- a/torchvision/transforms/v2/functional/_color.py
+++ b/torchvision/transforms/v2/functional/_color.py
@@ -79,14 +79,16 @@ def grayscale_to_rgb(inpt: torch.Tensor) -> torch.Tensor:
 @_register_kernel_internal(grayscale_to_rgb, torch.Tensor)
 @_register_kernel_internal(grayscale_to_rgb, tv_tensors.Image)
 def grayscale_to_rgb_image(image: torch.Tensor) -> torch.Tensor:
+    if image.shape[-3] >= 3:
+        # Image already has RGB channels. We don't need to do anything.
+        return image
     # rgb_to_grayscale can be used to add channels so we reuse that function.
     return _rgb_to_grayscale_image(image, num_output_channels=3, preserve_dtype=True)
 
 
 @_register_kernel_internal(grayscale_to_rgb, PIL.Image.Image)
 def grayscale_to_rgb_image_pil(image: PIL.Image.Image) -> PIL.Image.Image:
-    # to_grayscale can expand channels from 1 to 3 so we reuse that function.
-    return _FP.to_grayscale(image, num_output_channels=3)
+    return image.convert(mode="RGB")
 
 
 def _blend(image1: torch.Tensor, image2: torch.Tensor, ratio: float) -> torch.Tensor:

From dde638e641cd106c13c9ce9c2c5fa631495beb5d Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Mon, 12 Feb 2024 13:07:30 -0800
Subject: [PATCH 09/14] Removed support for v1 transform class

---
 torchvision/transforms/_functional_pil.py    |  7 -----
 torchvision/transforms/_functional_tensor.py | 13 ---------
 torchvision/transforms/functional.py         | 19 -------------
 torchvision/transforms/transforms.py         | 28 --------------------
 torchvision/transforms/v2/_color.py          |  2 --
 5 files changed, 69 deletions(-)

diff --git a/torchvision/transforms/_functional_pil.py b/torchvision/transforms/_functional_pil.py
index f82684fafb9..277848224ac 100644
--- a/torchvision/transforms/_functional_pil.py
+++ b/torchvision/transforms/_functional_pil.py
@@ -348,13 +348,6 @@ def to_grayscale(img: Image.Image, num_output_channels: int) -> Image.Image:
     return img
 
 
-def grayscale_to_rgb(img: Image.Image) -> Image.Image:
-    if not _is_pil_image(img):
-        raise TypeError(f"img should be PIL Image. Got {type(img)}")
-
-    return img.convert("RGB")
-
-
 @torch.jit.unused
 def invert(img: Image.Image) -> Image.Image:
     if not _is_pil_image(img):
diff --git a/torchvision/transforms/_functional_tensor.py b/torchvision/transforms/_functional_tensor.py
index 9bccdfe5d9f..88dc9ca21cc 100644
--- a/torchvision/transforms/_functional_tensor.py
+++ b/torchvision/transforms/_functional_tensor.py
@@ -166,19 +166,6 @@ def rgb_to_grayscale(img: Tensor, num_output_channels: int = 1) -> Tensor:
     return l_img
 
 
-def grayscale_to_rgb(img: Tensor) -> Tensor:
-    if img.ndim < 3:
-        raise TypeError(f"Input image tensor should have at least 3 dimensions, but found {img.ndim}")
-    _assert_channels(img, [1, 3])
-
-    if img.shape[-3] != 1:
-        raise ValueError("Expected last dimension of input image tensor to be 1, got {}".format(img.shape[-3]))
-
-    s = [-1] * len(img.shape)
-    s[-3] = 3
-    return img.expand(s)
-
-
 def adjust_brightness(img: Tensor, brightness_factor: float) -> Tensor:
     if brightness_factor < 0:
         raise ValueError(f"brightness_factor ({brightness_factor}) is not non-negative.")
diff --git a/torchvision/transforms/functional.py b/torchvision/transforms/functional.py
index c83bf018adc..698942c56af 100644
--- a/torchvision/transforms/functional.py
+++ b/torchvision/transforms/functional.py
@@ -1276,25 +1276,6 @@ def rgb_to_grayscale(img: Tensor, num_output_channels: int = 1) -> Tensor:
     return F_t.rgb_to_grayscale(img, num_output_channels)
 
 
-def grayscale_to_rgb(img: Tensor) -> Tensor:
-    """Converts grayscale image to 3-channel image by duplicating the value in all three channels.
-    If the image is torch Tensor, it is expected
-    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
-
-    Note:
-        Please, note that this method supports only grayscale images as input. For inputs in other color spaces,
-        please, consider using :meth:`~torchvision.transforms.functional.to_grayscale` with PIL Image.
-
-    Returns:
-        PIL Image or Tensor: RGB Image.
-    """
-    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
-        _log_api_usage_once(grayscale_to_rgb)
-    if not isinstance(img, torch.Tensor):
-        return F_pil.grayscale_to_rgb(img)
-    return F_t.grayscale_to_rgb(img)
-
-
 def erase(img: Tensor, i: int, j: int, h: int, w: int, v: Tensor, inplace: bool = False) -> Tensor:
     """Erase the input Tensor Image with given value.
     This transform does not support PIL Image.
diff --git a/torchvision/transforms/transforms.py b/torchvision/transforms/transforms.py
index 8ec3e3466af..2a6e0ce12c0 100644
--- a/torchvision/transforms/transforms.py
+++ b/torchvision/transforms/transforms.py
@@ -42,7 +42,6 @@
     "RandomRotation",
     "RandomAffine",
     "Grayscale",
-    "GrayscaleToRgb",
     "RandomGrayscale",
     "RandomPerspective",
     "RandomErasing",
@@ -1577,33 +1576,6 @@ def __repr__(self) -> str:
         return f"{self.__class__.__name__}(num_output_channels={self.num_output_channels})"
 
 
-class GrayscaleToRgb(torch.nn.Module):
-    """Converts grayscale image to 3 channel RGB image.
-    If the image is torch Tensor, it is expected
-    to have [..., 1, H, W] shape, where ... means an arbitrary number of leading dimensions
-
-    Returns:
-        PIL Image: RGB version of the input image.
-    """
-
-    def __init__(self):
-        super().__init__()
-        _log_api_usage_once(self)
-
-    def forward(self, img):
-        """
-        Args:
-            img (PIL Image or Tensor): Image to be converted to grayscale.
-
-        Returns:
-            PIL Image or Tensor: Grayscaled image.
-        """
-        return F.grayscale_to_rgb(img)
-
-    def __repr__(self) -> str:
-        return f"{self.__class__.__name__}()"
-
-
 class RandomGrayscale(torch.nn.Module):
     """Randomly convert image to grayscale with a probability of p (default 0.1).
     If the image is torch Tensor, it is expected
diff --git a/torchvision/transforms/v2/_color.py b/torchvision/transforms/v2/_color.py
index 39605f46ed7..0979afd5837 100644
--- a/torchvision/transforms/v2/_color.py
+++ b/torchvision/transforms/v2/_color.py
@@ -61,8 +61,6 @@ class GrayscaleToRgb(Transform):
     to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions
     """
 
-    _v1_transform_cls = _transforms.GrayscaleToRgb
-
     def __init__(self):
         super().__init__()
 

From dd01a21e93f08f4ee14ead912cdcfc8361604d0b Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Mon, 4 Mar 2024 12:27:26 -0800
Subject: [PATCH 10/14] Accepted suggestion that simplifies test code

---
 test/test_transforms_v2.py | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)

diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
index 2af4dd9d739..42d5f87045e 100644
--- a/test/test_transforms_v2.py
+++ b/test/test_transforms_v2.py
@@ -5017,15 +5017,7 @@ def test_expanded_channels_are_not_views_into_the_same_underlying_tensor(self):
     def test_rgb_image_is_unchanged(self):
         image = make_image(dtype=torch.uint8, device="cpu", color_space="RGB")
         assert_equal(image.shape[-3], 3)
-        image[0][0][0] = 0
-        image[1][0][0] = 100
-        image[2][0][0] = 200
-        output_image = F.grayscale_to_rgb(image)
-        assert output_image[0][0][0] == 0
-        assert output_image[1][0][0] == 100
-        assert output_image[2][0][0] == 200
-        print(image)
-        print(output_image)
+        assert_equal(F.grayscale_to_rgb(image), image)
 
 
 class TestRandomZoomOut:

From 02526afc628f40f4100593450b2228e07cc9901f Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Fri, 15 Mar 2024 08:23:09 -0700
Subject: [PATCH 11/14] Changed names

---
 docs/source/transforms.rst            | 4 +++-
 test/test_transforms_v2.py            | 4 ++--
 torchvision/transforms/v2/__init__.py | 2 +-
 torchvision/transforms/v2/_color.py   | 2 +-
 4 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/docs/source/transforms.rst b/docs/source/transforms.rst
index 056d1589e84..b52914a373c 100644
--- a/docs/source/transforms.rst
+++ b/docs/source/transforms.rst
@@ -347,6 +347,7 @@ Color
     v2.RandomChannelPermutation
     v2.RandomPhotometricDistort
     v2.Grayscale
+    v2.RGB
     v2.RandomGrayscale
     v2.GaussianBlur
     v2.RandomInvert
@@ -364,6 +365,7 @@ Functionals
 
     v2.functional.permute_channels
     v2.functional.rgb_to_grayscale
+    v2.functional.grayscale_to_rgb
     v2.functional.to_grayscale
     v2.functional.gaussian_blur
     v2.functional.invert
@@ -583,7 +585,7 @@ Conversion
     while performing the conversion, while some may not do any scaling. By
     scaling, we mean e.g. that a ``uint8`` -> ``float32`` would map the [0,
     255] range into [0, 1] (and vice-versa). See :ref:`range_and_dtype`.
-    
+
 .. autosummary::
     :toctree: generated/
     :template: class.rst
diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
index 1a247625560..1dd3eae2526 100644
--- a/test/test_transforms_v2.py
+++ b/test/test_transforms_v2.py
@@ -5012,9 +5012,9 @@ def test_functional_signature(self, kernel, input_type):
 
     @pytest.mark.parametrize("make_input", [make_image_tensor, make_image_pil, make_image])
     def test_transform(self, make_input):
-        check_transform(transforms.GrayscaleToRgb(), make_input(color_space="GRAY"))
+        check_transform(transforms.RGB(), make_input(color_space="GRAY"))
 
-    @pytest.mark.parametrize("fn", [F.grayscale_to_rgb, transform_cls_to_functional(transforms.GrayscaleToRgb)])
+    @pytest.mark.parametrize("fn", [F.grayscale_to_rgb, transform_cls_to_functional(transforms.RGB)])
     def test_image_correctness(self, fn):
         image = make_image(dtype=torch.uint8, device="cpu", color_space="GRAY")
 
diff --git a/torchvision/transforms/v2/__init__.py b/torchvision/transforms/v2/__init__.py
index 2cb3ea3eeae..61972994639 100644
--- a/torchvision/transforms/v2/__init__.py
+++ b/torchvision/transforms/v2/__init__.py
@@ -9,7 +9,7 @@
 from ._color import (
     ColorJitter,
     Grayscale,
-    GrayscaleToRgb,
+    RGB,
     RandomAdjustSharpness,
     RandomAutocontrast,
     RandomChannelPermutation,
diff --git a/torchvision/transforms/v2/_color.py b/torchvision/transforms/v2/_color.py
index 0979afd5837..ab049b6bf9e 100644
--- a/torchvision/transforms/v2/_color.py
+++ b/torchvision/transforms/v2/_color.py
@@ -54,7 +54,7 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
         return self._call_kernel(F.rgb_to_grayscale, inpt, num_output_channels=params["num_input_channels"])
 
 
-class GrayscaleToRgb(Transform):
+class RGB(Transform):
     """Converts grayscale images to RGB images.
 
     If the input is a :class:`torch.Tensor`, it is expected

From d24b3222078f1d8ea8d15848cd94d7fe2e49e8b3 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Fri, 15 Mar 2024 08:24:34 -0700
Subject: [PATCH 12/14] Tweaked documentation

---
 torchvision/transforms/v2/_color.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/torchvision/transforms/v2/_color.py b/torchvision/transforms/v2/_color.py
index ab049b6bf9e..8f334105911 100644
--- a/torchvision/transforms/v2/_color.py
+++ b/torchvision/transforms/v2/_color.py
@@ -55,7 +55,7 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
 
 
 class RGB(Transform):
-    """Converts grayscale images to RGB images.
+    """Converts images or videos to RGB (if they are already not RGB)
 
     If the input is a :class:`torch.Tensor`, it is expected
     to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions

From 7331023cbb35b4a716087e51564fce5d7adff962 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Fri, 15 Mar 2024 08:32:42 -0700
Subject: [PATCH 13/14] Sorted imports

---
 torchvision/transforms/v2/__init__.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/torchvision/transforms/v2/__init__.py b/torchvision/transforms/v2/__init__.py
index 61972994639..fea39d3cf20 100644
--- a/torchvision/transforms/v2/__init__.py
+++ b/torchvision/transforms/v2/__init__.py
@@ -9,7 +9,6 @@
 from ._color import (
     ColorJitter,
     Grayscale,
-    RGB,
     RandomAdjustSharpness,
     RandomAutocontrast,
     RandomChannelPermutation,
@@ -19,6 +18,7 @@
     RandomPhotometricDistort,
     RandomPosterize,
     RandomSolarize,
+    RGB,
 )
 from ._container import Compose, RandomApply, RandomChoice, RandomOrder
 from ._geometry import (

From b2daebd998c9d5ca2a47d054c3ba9133afa883c5 Mon Sep 17 00:00:00 2001
From: Ahmad Sharif <ahmads@fb.com>
Date: Fri, 15 Mar 2024 08:43:54 -0700
Subject: [PATCH 14/14] Adhere to PEP0257

---
 torchvision/transforms/v2/_color.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/torchvision/transforms/v2/_color.py b/torchvision/transforms/v2/_color.py
index 8f334105911..49b4a8d8b10 100644
--- a/torchvision/transforms/v2/_color.py
+++ b/torchvision/transforms/v2/_color.py
@@ -55,7 +55,7 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
 
 
 class RGB(Transform):
-    """Converts images or videos to RGB (if they are already not RGB)
+    """Convert images or videos to RGB (if they are already not RGB).
 
     If the input is a :class:`torch.Tensor`, it is expected
     to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions