📚 The doc issue
Following URLs in the ET doc page have outdated versions in the URL:
- Using AAR file directly under Using ET in Android
- Android’s Qualcomm backend section
-
XNNPACK’s Partitioner API’s constructor, ../config/init.py, ConfigPrecisionType links have v0.6
-
Blob links in https://docs.pytorch.org/executorch/stable/backends-qualcomm.html
-
Arm Ethos-U Troubleshooting: bundled.io link is v0.4
-
Example python script to convert a custom PyTorch model using NXP eIQ has v1.0
-
Custom delegation options in NXP eIQ’s Partitioner API has v1.0
-
Node convertors under operation support in NXP eIQ has v1.0
-
Aot_neutron_compile.py’s readme in Model prep step for NXP eiQ Neutron tutorial has v1.0
-
Custom Compiler Passes links to a particular blob than main
-
Greedy Algorithm’s implementation in Custom Memory Plans under Compiler & IR has a specific blob instead of main
-
Bundled Program (since just example, not mandatory to change)22. Example usage of get_program_data under Runtime Stage for links to v1.0
- API Life Cycle & Deprecation Policy (since just example, not mandatory to change)
- Example usage for using .. warning:: in the docstrings of deprecated and experimental APIs
- Example usage to use the ET_DEPRECATED annotation macro.
- Example usage to Start Doxygen comments with DEPRECATED:
Suggest a potential alternative/fix
link to /stable version for docs and /main/ instead of particular blob in the case of GitHub URLS to ensure the most updated URL is always linked
cc @mergennachin @AlannaBurke
📚 The doc issue
Following URLs in the ET doc page have outdated versions in the URL:
XNNPACK’s Partitioner API’s constructor, ../config/init.py, ConfigPrecisionType links have v0.6
Blob links in https://docs.pytorch.org/executorch/stable/backends-qualcomm.html
Arm Ethos-U Troubleshooting: bundled.io link is v0.4
Example python script to convert a custom PyTorch model using NXP eIQ has v1.0
Custom delegation options in NXP eIQ’s Partitioner API has v1.0
Node convertors under operation support in NXP eIQ has v1.0
Aot_neutron_compile.py’s readme in Model prep step for NXP eiQ Neutron tutorial has v1.0
Custom Compiler Passes links to a particular blob than main
Greedy Algorithm’s implementation in Custom Memory Plans under Compiler & IR has a specific blob instead of main
Bundled Program (since just example, not mandatory to change)22. Example usage of get_program_data under Runtime Stage for links to v1.0
Suggest a potential alternative/fix
link to /stable version for docs and /main/ instead of particular blob in the case of GitHub URLS to ensure the most updated URL is always linked
cc @mergennachin @AlannaBurke