Fix memory leak in decode_webp

[NicolasHug]

Closes #8710

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/8712

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 16 New Failures, 1 Unrelated Failure

As of commit 14324f6 with merge base 868a3b4 ():

NEW FAILURES - The following jobs have failed:

• Build Aarch64 Linux Wheels / pytorch/vision / upload / wheel-py3_9-cpu-aarch64 (gh)
  ##[error]Unable to find any artifacts for the associated workflow
• Build Aarch64 Linux Wheels / pytorch/vision / upload / wheel-py3_9-cuda-aarch64cuda-aarch64 (gh)
  ##[error]Unable to find any artifacts for the associated workflow
• Build Aarch64 Linux Wheels / pytorch/vision / wheel-py3_9-cpu-aarch64 (gh)
  /__w/_temp/65554edc-ff95-401d-a6ea-75ca731f6ef3.sh: line 2: conda: command not found
• Build Aarch64 Linux Wheels / pytorch/vision / wheel-py3_9-cuda-aarch64cuda-aarch64 (gh)
  /__w/_temp/b1b58b47-a252-420a-92e1-7f5fdde6b459.sh: line 2: conda: command not found
• Build Linux Wheels / pytorch/vision / upload / manywheel-py3_9-cpu (gh)
  ##[error]Unable to find an artifact with the name: pytorch_vision__3.9_cpu_x86_64
• Build Linux Wheels / pytorch/vision / upload / manywheel-py3_9-cuda11_8 (gh)
  ##[error]Unable to find an artifact with the name: pytorch_vision__3.9_cu118_x86_64
• Build Linux Wheels / pytorch/vision / upload / manywheel-py3_9-cuda12_1 (gh)
  ##[error]Unable to find an artifact with the name: pytorch_vision__3.9_cu121_x86_64
• Build Linux Wheels / pytorch/vision / upload / manywheel-py3_9-cuda12_4 (gh)
  ##[error]Unable to find an artifact with the name: pytorch_vision__3.9_cu124_x86_64
• Build Linux Wheels / pytorch/vision / upload / manywheel-py3_9-rocm6_1 (gh)
  ##[error]Unable to find an artifact with the name: pytorch_vision__3.9_rocm6.1_x86_64
• Build Linux Wheels / pytorch/vision / upload / manywheel-py3_9-rocm6_2 (gh)
  ##[error]Unable to find an artifact with the name: pytorch_vision__3.9_rocm6.2_x86_64
• Docs / build / linux-job (gh)
  RuntimeError: Command docker exec -t 9376c824e66f72f71f0113e77287a7c2af905957529b8f179ba749edb00c2417 /exec failed with exit code 2
• Lint / c-source / linux-job (gh)
  RuntimeError: Command docker exec -t 34267766f58c11a157ec16b735ff2e02cd007d2ed4f41fcf79a155d2a1e9cd8d /exec failed with exit code 1
• Tests / unittests-linux (3.10, linux.12xlarge, cpu) / linux-job (gh)
  test/test_ops.py::TestDeformConv::test_backward[0-True-cpu]
• Tests / unittests-linux (3.11, linux.12xlarge, cpu) / linux-job (gh)
  test/test_ops.py::TestDeformConv::test_backward[0-True-cpu]
• Tests / unittests-linux (3.12, linux.12xlarge, cpu) / linux-job (gh)
  test/test_ops.py::TestDeformConv::test_backward[0-True-cpu]
• Tests / unittests-linux (3.9, linux.12xlarge, cpu) / linux-job (gh)
  test/test_ops.py::TestDeformConv::test_backward[0-True-cpu]

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• Tests / unittests-windows (3.9, windows.g5.4xlarge.nvidia.gpu, cuda, 11.8) / windows-job (gh) (trunk failure)
  AttributeError: '_OpNamespace' 'image' object has no attribute '_jpeg_version'

This comment was automatically generated by Dr. CI and updates every 15 minutes.