Update cuda-bindings minimum version from 13.3.0 to 13.1.0

Address review comment: cudaGraphNodeGetToolsId API was introduced in
CUDA 13.1.0, not 13.3.0. Update requirements and tutorial documentation
to reflect the correct minimum version.

Verified via CUDA documentation that the API first appeared in 13.1.0.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

Author: yushangdi

.ci/docker/requirements.txt

@@ -26,7 +26,7 @@ fastapi
 matplotlib
 librosa
 torch==2.12
-cuda-bindings>=13.3.0  # Required for CUDA graph annotations tutorial
+cuda-bindings>=13.1.0  # Required for CUDA graph annotations tutorial
 torchvision
 torchdata
 networkx

advanced_source/cuda_graph_annotations_tutorial.py

@@ -23,7 +23,7 @@
        * PyTorch 2.12+
        * CUDA-capable GPU
        * Driver/CUDA-compat >= 13.1 for annotation support
-       * cuda-bindings >= 13.3.0
+       * cuda-bindings >= 13.1.0
 
 CUDA graphs are a powerful optimization technique that can significantly reduce
 kernel launch overhead by capturing and replaying sequences of CUDA operations.
@@ -81,10 +81,10 @@
 # - PyTorch 2.12+
 # - A CUDA GPU
 # - Driver/CUDA-compat >= 13.1 for annotation support
-# - The ``cuda-bindings`` package >= 13.3.0 (``pip install cuda-python``)
+# - The ``cuda-bindings`` package >= 13.1.0 (``pip install cuda-python``)
 #
 # The cuda-bindings package provides the Python bindings for CUDA runtime APIs.
-# Version 13.3.0+ is required for the ``cudaGraphNodeGetToolsId`` API that
+# Version 13.1.0+ is required for the ``cudaGraphNodeGetToolsId`` API that
 # enables kernel annotations. If you have an older version, the tutorial will
 # run but annotations will be disabled with a warning message explaining how
 # to upgrade.