Add option to initialize optical flow in RAFT

[ChristophReich1996]

Hi, this pull request adds support to initialize the optical flow to be refined by the RAFT decoder. The current implementation only supports starting from zero optical flow. However, the original RAFT paper also proposes initializing the optical flow, e.g., with the flow prediction of previous frames. This can lead to a faster inference time since fewer iterative refinements are typically needed. This pull request adapts from the original RAFT implementation.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/8521

• 📄 Preview Python docs built from this PR

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

❌ 18 New Failures

As of commit ac5eae6 with merge base 8f9d810 ():

NEW FAILURES - The following jobs have failed:

• Build Linux Conda / pytorch/vision / upload / conda-py3_8-cpu (gh)
  Unable to find any artifacts for the associated workflow
• Build Linux Conda / pytorch/vision / upload / conda-py3_8-cuda11_8 (gh)
  Unable to find any artifacts for the associated workflow
• Build Linux Conda / pytorch/vision / upload / conda-py3_8-cuda12_1 (gh)
  Unable to find any artifacts for the associated workflow
• Build Linux Conda / pytorch/vision / upload / conda-py3_8-cuda12_4 (gh)
  Unable to find any artifacts for the associated workflow
• Build Windows Conda / pytorch/vision / upload / conda-py3_8-cpu (gh)
  Unable to find any artifacts for the associated workflow
• Build Windows Conda / pytorch/vision / upload / conda-py3_8-cuda11_8 (gh)
  Unable to find any artifacts for the associated workflow
• Build Windows Conda / pytorch/vision / upload / conda-py3_8-cuda12_1 (gh)
  Unable to find any artifacts for the associated workflow
• Build Windows Conda / pytorch/vision / upload / conda-py3_8-cuda12_4 (gh)
  Unable to find any artifacts for the associated workflow
• Lint / python-source-and-configs / linux-job (gh)
  RuntimeError: Command docker exec -t c4c1031b555e465881832a6ede8ef9fa0a809bb8993a0aa8ac3be6bdceef7036 /exec failed with exit code 1
• Lint / python-types / linux-job (gh)
  RuntimeError: Command docker exec -t 35ef998122bb36af37fdf9cda2a6d7991fe60d4e4b0100305b2b9b014076dc80 /exec failed with exit code 1
• Tests / unittests-linux (3.10, linux.12xlarge, cpu) / linux-job (gh)
  test/test_io.py::TestVideo::test_write_video_with_audio[cuda]
• Tests / unittests-linux (3.12, linux.12xlarge, cpu) / linux-job (gh)
  test/test_io.py::TestVideo::test_write_video_with_audio[cuda]
• Tests / unittests-linux (3.8, linux.12xlarge, cpu) / linux-job (gh)
  test/test_io.py::TestVideo::test_write_video_with_audio[cuda]
• Tests / unittests-linux (3.9, linux.12xlarge, cpu) / linux-job (gh)
  test/test_io.py::TestVideo::test_write_video_with_audio[cuda]
• Tests / unittests-macos (3.10, macos-m1-stable) / macos-job (gh)
  test/test_io.py::TestVideo::test_write_video_with_audio[cuda]
• Tests / unittests-macos (3.12, macos-m1-stable) / macos-job (gh)
  test/test_io.py::TestVideo::test_write_video_with_audio[cuda]
• Tests / unittests-macos (3.8, macos-m1-stable) / macos-job (gh)
  test/test_io.py::TestVideo::test_write_video_with_audio[cuda]
• Tests / unittests-macos (3.9, macos-m1-stable) / macos-job (gh)
  test/test_ops.py::TestRoIAlign::test_autocast_cpu[rois_dtype1-x_dtype1-False-True]

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

[NicolasHug]

Thanks for the PR @ChristophReich1996 . This seems reasonable to merge. Can you please remove all unrelated changes that were made to the code formatting? Thanks!