Add warning in docs and Cpp code for nvjpeg leak on CUDA < 11.6 (#5482)

torchvision/csrc/io/image/cuda/decode_jpeg_cuda.cpp

@@ -47,6 +47,27 @@ torch::Tensor decode_jpeg_cuda(
 
   TORCH_CHECK(device.is_cuda(), "Expected a cuda device")
 
+  int major_version;
+  int minor_version;
+  nvjpegStatus_t get_major_property_status =
+      nvjpegGetProperty(MAJOR_VERSION, &major_version);
+  nvjpegStatus_t get_minor_property_status =
+      nvjpegGetProperty(MINOR_VERSION, &minor_version);
+
+  TORCH_CHECK(
+      get_major_property_status == NVJPEG_STATUS_SUCCESS,
+      "nvjpegGetProperty failed: ",
+      get_major_property_status);
+  TORCH_CHECK(
+      get_minor_property_status == NVJPEG_STATUS_SUCCESS,
+      "nvjpegGetProperty failed: ",
+      get_minor_property_status);
+  if ((major_version < 11) || ((major_version == 11) && (minor_version < 6))) {
+    TORCH_WARN_ONCE(
+        "There is a memory leak issue in the nvjpeg library for CUDA versions < 11.6. "
+        "Make sure to rely on CUDA 11.6 or above before using decode_jpeg(..., device='cuda').");
+  }
+
   at::cuda::CUDAGuard device_guard(device);
 
   // Create global nvJPEG handle

torchvision/io/image.py

@@ -145,6 +145,10 @@ def decode_jpeg(
             with `nvjpeg <https://developer.nvidia.com/nvjpeg>`_. This is only
             supported for CUDA version >= 10.1
 
+            .. warning::
+                There is a memory leak in the nvjpeg library for CUDA versions < 11.6.
+                Make sure to rely on CUDA 11.6 or above before using ``device="cuda"``.
+
     Returns:
         output (Tensor[image_channels, image_height, image_width])
     """