Add torchvision

Florence2 model preprocessor needs it to accelerate process

pdm.lock

@@ -5,7 +5,7 @@
 groups = ["default"]
 strategy = ["inherit_metadata"]
 lock_version = "4.5.0"
-content_hash = "sha256:347bfe6bad5a51324dd0fad3ab415f5a38c826d0ac0fe6a1b6d3e873a6b29168"
+content_hash = "sha256:8f5953bb60ddf9d7b083a2f10f3171bdd4227baf682b30b520b6391ea160f7a5"
 
 [[metadata.targets]]
 requires_python = ">=3.10,<3.13"
@@ -1432,6 +1432,32 @@ files = [
     {file = "torch-2.9.1-cp312-none-macosx_11_0_arm64.whl", hash = "sha256:c0d25d1d8e531b8343bea0ed811d5d528958f1dcbd37e7245bc686273177ad7e"},
 ]
 
+[[package]]
+name = "torchvision"
+version = "0.24.1"
+requires_python = ">=3.10"
+summary = "image and video datasets and models for torch deep learning"
+groups = ["default"]
+dependencies = [
+    "numpy",
+    "pillow!=8.3.*,>=5.3.0",
+    "torch==2.9.1",
+]
+files = [
+    {file = "torchvision-0.24.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:ded5e625788572e4e1c4d155d1bbc48805c113794100d70e19c76e39e4d53465"},
+    {file = "torchvision-0.24.1-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:54ed17c3d30e718e08d8da3fd5b30ea44b0311317e55647cb97077a29ecbc25b"},
+    {file = "torchvision-0.24.1-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:f476da4e085b7307aaab6f540219617d46d5926aeda24be33e1359771c83778f"},
+    {file = "torchvision-0.24.1-cp310-cp310-win_amd64.whl", hash = "sha256:fbdbdae5e540b868a681240b7dbd6473986c862445ee8a138680a6a97d6c34ff"},
+    {file = "torchvision-0.24.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:056c525dc875f18fe8e9c27079ada166a7b2755cea5a2199b0bc7f1f8364e600"},
+    {file = "torchvision-0.24.1-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:1e39619de698e2821d71976c92c8a9e50cdfd1e993507dfb340f2688bfdd8283"},
+    {file = "torchvision-0.24.1-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:a0f106663e60332aa4fcb1ca2159ef8c3f2ed266b0e6df88de261048a840e0df"},
+    {file = "torchvision-0.24.1-cp311-cp311-win_amd64.whl", hash = "sha256:a9308cdd37d8a42e14a3e7fd9d271830c7fecb150dd929b642f3c1460514599a"},
+    {file = "torchvision-0.24.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:e48bf6a8ec95872eb45763f06499f87bd2fb246b9b96cb00aae260fda2f96193"},
+    {file = "torchvision-0.24.1-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:7fb7590c737ebe3e1c077ad60c0e5e2e56bb26e7bccc3b9d04dbfc34fd09f050"},
+    {file = "torchvision-0.24.1-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:66a98471fc18cad9064123106d810a75f57f0838eee20edc56233fd8484b0cc7"},
+    {file = "torchvision-0.24.1-cp312-cp312-win_amd64.whl", hash = "sha256:4aa6cb806eb8541e92c9b313e96192c6b826e9eb0042720e2fa250d021079952"},
+]
+
 [[package]]
 name = "tqdm"
 version = "4.67.1"

pyproject.toml

@@ -10,6 +10,7 @@ dependencies = [
   "pillow>=12.0.0",
   "spaces>=0.44.0",
   "networkx>=3.0,<3.6",
+  "torchvision>=0.24.1",
 ]
 requires-python = ">=3.10,<3.13"
 readme = "README.md"

requirements.txt

@@ -74,6 +74,7 @@ sympy==1.14.0
 tokenizers==0.22.1
 tomlkit==0.13.3
 torch>=2.4.0
+torchvision
 tqdm==4.67.1
 transformers==4.57.3
 triton==3.5.1; platform_system == "Linux" and platform_machine == "x86_64"