2025-02-27 nightly release (492a6bc)

Author: pytorchbot

.github/workflows/docs.yaml

@@ -75,9 +75,6 @@ jobs:
         uses: pytorch/test-infra/.github/actions/setup-miniconda@main
         with:
           python-version: ${{ matrix.python-version }}
-          #
-          # For some reason nvidia::libnpp=12.4 doesn't install but nvidia/label/cuda-12.4.0::libnpp does.
-          # So we use the latter convention for libnpp.
           # We install conda packages at the start because otherwise conda may have conflicts with dependencies.
           default-packages: "nvidia/label/cuda-${{ matrix.cuda-version }}.0::libnpp nvidia::cuda-nvrtc=${{ matrix.cuda-version }} nvidia::cuda-toolkit=${{ matrix.cuda-version }} nvidia::cuda-cudart=${{ matrix.cuda-version }} nvidia::cuda-driver-dev=${{ matrix.cuda-version }} conda-forge::ffmpeg=${{ matrix.ffmpeg-version-for-tests }}"
       - name: Check env

.github/workflows/linux_cuda_wheel.yaml

@@ -66,12 +66,9 @@ jobs:
           # For the actual release we should add that label and change this to
           # include more python versions.
         python-version: ['3.9']
-        cuda-version: ['11.8', '12.4', '12.6']
+        cuda-version: ['11.8', '12.6']
         # TODO: put back ffmpeg 5 https://github.com/pytorch/torchcodec/issues/325
         ffmpeg-version-for-tests: ['4.4.2', '6', '7']
-        exclude:
-        - cuda-version: '12.4'  # TODO put this back it fails with infra issue.
-          ffmpeg-version-for-tests: '4.4.2'
 
     container:
       image: "pytorch/manylinux2_28-builder:cuda${{ matrix.cuda-version }}"
@@ -92,9 +89,6 @@ jobs:
         uses: pytorch/test-infra/.github/actions/setup-miniconda@main
         with:
           python-version: ${{ matrix.python-version }}
-          #
-          # For some reason nvidia::libnpp=12.4 doesn't install but nvidia/label/cuda-12.4.0::libnpp does.
-          # So we use the latter convention for libnpp.
           # We install conda packages at the start because otherwise conda may have conflicts with dependencies.
           default-packages: "nvidia/label/cuda-${{ matrix.cuda-version }}.0::libnpp nvidia::cuda-nvrtc=${{ matrix.cuda-version }} nvidia::cuda-toolkit=${{ matrix.cuda-version }} nvidia::cuda-cudart=${{ matrix.cuda-version }} nvidia::cuda-driver-dev=${{ matrix.cuda-version }} conda-forge::ffmpeg=${{ matrix.ffmpeg-version-for-tests }}"
       - name: Check env

README.md

@@ -146,14 +146,12 @@ First, make sure you have a GPU that has NVDEC hardware that can decode the
 format you want. Refer to Nvidia's GPU support matrix for more details
 [here](https://developer.nvidia.com/video-encode-and-decode-gpu-support-matrix-new).
 
-1. Install CUDA Toolkit. Pytorch and TorchCodec supports CUDA Toolkit
-   versions 11.8, 12.1 or 12.4. In particular TorchCodec depends on
-   CUDA libraries libnpp and libnvrtc (which are part of CUDA Toolkit).
+1. Install Pytorch corresponding to your CUDA Toolkit using the
+   [official instructions](https://pytorch.org/get-started/locally/). You'll
+   need the `libnpp` and `libnvrtc` CUDA libraries, which are usually part of
+   the CUDA Toolkit.
 
-2. Install Pytorch that corresponds to your CUDA Toolkit version using the
-   [official instructions](https://pytorch.org/get-started/locally/).
-
-3. Install or compile FFmpeg with NVDEC support.
+2. Install or compile FFmpeg with NVDEC support.
    TorchCodec with CUDA should work with FFmpeg versions in [5, 7].
 
    If FFmpeg is not already installed, or you need a more recent version, an
@@ -184,13 +182,13 @@ format you want. Refer to Nvidia's GPU support matrix for more details
    ffmpeg -hwaccel cuda -hwaccel_output_format cuda -i test/resources/nasa_13013.mp4 -f null -
    ```
 
-4. Install TorchCodec by passing in an `--index-url` parameter that corresponds
+3. Install TorchCodec by passing in an `--index-url` parameter that corresponds
    to your CUDA Toolkit version, example:
 
    ```bash
-   # This corresponds to CUDA Toolkit version 12.4. It should be the same one
+   # This corresponds to CUDA Toolkit version 12.6. It should be the same one
    # you used when you installed PyTorch (If you installed PyTorch with pip).
-   pip install torchcodec --index-url=https://download.pytorch.org/whl/cu124
+   pip install torchcodec --index-url=https://download.pytorch.org/whl/cu126
    ```
 
    Note that without passing in the `--index-url` parameter, `pip` installs

src/torchcodec/decoders/_core/video_decoder_ops.py

@@ -185,7 +185,6 @@ def get_frame_at_pts_abstract(
 def get_frames_by_pts_abstract(
     decoder: torch.Tensor,
     *,
-    stream_index: int,
     timestamps: List[float],
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
     image_size = [get_ctx().new_dynamic_size() for _ in range(4)]
@@ -198,7 +197,7 @@ def get_frames_by_pts_abstract(
 
 @register_fake("torchcodec_ns::get_frame_at_index")
 def get_frame_at_index_abstract(
-    decoder: torch.Tensor, *, stream_index: int, frame_index: int
+    decoder: torch.Tensor, *, frame_index: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
     image_size = [get_ctx().new_dynamic_size() for _ in range(3)]
     return (
@@ -212,7 +211,6 @@ def get_frame_at_index_abstract(
 def get_frames_at_indices_abstract(
     decoder: torch.Tensor,
     *,
-    stream_index: int,
     frame_indices: List[int],
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
     image_size = [get_ctx().new_dynamic_size() for _ in range(4)]
@@ -227,7 +225,6 @@ def get_frames_at_indices_abstract(
 def get_frames_in_range_abstract(
     decoder: torch.Tensor,
     *,
-    stream_index: int,
     start: int,
     stop: int,
     step: Optional[int] = None,
@@ -244,7 +241,6 @@ def get_frames_in_range_abstract(
 def get_frames_by_pts_in_range_abstract(
     decoder: torch.Tensor,
     *,
-    stream_index: int,
     start_seconds: float,
     stop_seconds: float,
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -257,9 +253,7 @@ def get_frames_by_pts_in_range_abstract(
 
 
 @register_fake("torchcodec_ns::_get_key_frame_indices")
-def get_key_frame_indices_abstract(
-    decoder: torch.Tensor, *, stream_index: int
-) -> torch.Tensor:
+def get_key_frame_indices_abstract(decoder: torch.Tensor) -> torch.Tensor:
     return torch.empty([], dtype=torch.int)
 
 
@@ -282,7 +276,6 @@ def get_stream_json_metadata_abstract(decoder: torch.Tensor, stream_idx: int) ->
 def _test_frame_pts_equality_abstract(
     decoder: torch.Tensor,
     *,
-    stream_index: int,
     frame_index: int,
     pts_seconds_to_test: float,
 ) -> bool:

test/decoders/test_video_decoder.py

@@ -4,6 +4,8 @@
 # This source code is licensed under the BSD-style license found in the
 # LICENSE file in the root directory of this source tree.
 
+import contextlib
+
 import numpy
 import pytest
 import torch
@@ -874,6 +876,38 @@ def test_get_key_frame_indices(self, device):
             key_frame_indices, h265_reference_key_frame_indices, atol=0, rtol=0
         )
 
+    @pytest.mark.parametrize("device", cpu_and_cuda())
+    def test_compile(self, device):
+        decoder = VideoDecoder(NASA_VIDEO.path, device=device)
 
-if __name__ == "__main__":
-    pytest.main()
+        @contextlib.contextmanager
+        def restore_capture_scalar_outputs():
+            try:
+                original = torch._dynamo.config.capture_scalar_outputs
+                yield
+            finally:
+                torch._dynamo.config.capture_scalar_outputs = original
+
+        # TODO: We get a graph break because we call Tensor.item() to turn the
+        # tensors in FrameBatch into scalars. When we work on compilation and exportability,
+        # we should investigate.
+        with restore_capture_scalar_outputs():
+            torch._dynamo.config.capture_scalar_outputs = True
+
+            @torch.compile(fullgraph=True, backend="eager")
+            def get_some_frames(decoder):
+                frames = []
+                frames.append(decoder.get_frame_at(1))
+                frames.append(decoder.get_frame_at(3))
+                frames.append(decoder.get_frame_at(5))
+                return frames
+
+            frames = get_some_frames(decoder)
+
+            ref_frame1 = NASA_VIDEO.get_frame_data_by_index(1).to(device)
+            ref_frame3 = NASA_VIDEO.get_frame_data_by_index(3).to(device)
+            ref_frame5 = NASA_VIDEO.get_frame_data_by_index(5).to(device)
+
+            assert_frames_equal(ref_frame1, frames[0].data)
+            assert_frames_equal(ref_frame3, frames[1].data)
+            assert_frames_equal(ref_frame5, frames[2].data)

test/decoders/test_video_decoder_ops.py

@@ -9,7 +9,6 @@
 os.environ["TORCH_LOGS"] = "output_code"
 import json
 import subprocess
-from typing import Tuple
 
 import numpy as np
 import pytest
@@ -48,20 +47,6 @@
 INDEX_OF_FRAME_AT_6_SECONDS = 180
 
 
-class ReferenceDecoder:
-    def __init__(self, device="cpu"):
-        self.decoder: torch.Tensor = create_from_file(str(NASA_VIDEO.path))
-        add_video_stream(self.decoder, device=device)
-
-    def get_next_frame(self) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
-        assert self.decoder is not None
-        return get_next_frame(self.decoder)
-
-    def seek(self, pts: float):
-        assert self.decoder is not None
-        seek_to_pts(self.decoder, pts)
-
-
 class TestOps:
     @pytest.mark.parametrize("device", cpu_and_cuda())
     def test_seek_and_next(self, device):
@@ -352,27 +337,6 @@ def get_frame1_and_frame_time6(decoder):
         assert_frames_equal(frame0, reference_frame0.to(device))
         assert_frames_equal(frame_time6, reference_frame_time6.to(device))
 
-    @pytest.mark.parametrize("device", cpu_and_cuda())
-    def test_class_based_compile_seek_and_next(self, device):
-        # TODO_OPEN_ISSUE Scott (T180277797): Ditto as above.
-        @torch.compile(fullgraph=True, backend="eager")
-        def class_based_get_frame1_and_frame_time6(
-            decoder: ReferenceDecoder,
-        ) -> Tuple[torch.Tensor, torch.Tensor]:
-            frame0, _, _ = decoder.get_next_frame()
-            decoder.seek(6.0)
-            frame_time6, _, _ = decoder.get_next_frame()
-            return frame0, frame_time6
-
-        decoder = ReferenceDecoder(device=device)
-        frame0, frame_time6 = class_based_get_frame1_and_frame_time6(decoder)
-        reference_frame0 = NASA_VIDEO.get_frame_data_by_index(0)
-        reference_frame_time6 = NASA_VIDEO.get_frame_data_by_index(
-            INDEX_OF_FRAME_AT_6_SECONDS
-        )
-        assert_frames_equal(frame0, reference_frame0.to(device))
-        assert_frames_equal(frame_time6, reference_frame_time6.to(device))
-
     @pytest.mark.parametrize("device", cpu_and_cuda())
     @pytest.mark.parametrize("create_from", ("file", "tensor", "bytes"))
     def test_create_decoder(self, create_from, device):