[DO NOT UNPIN] ORT 1.21.0 Release Candidates available for testing

[MaanavD]

ORT 1.21 will be released soon, and release candidates are now available for testing here! Please respond to report any bugs you encounter!

Release branch: rel-1.21.0
Release manager: @amarin16

Python Whls	Nuget	NPM	Maven (Java)
CPU: 1.21.0.dev20250228001	CPU: 1.21.0-dev-20250228	onnxruntime-node: 1.21.0-dev.20250228	CPU: 1.21.0-rc1
GPU: 1.21.0.dev20250228002	GPU (CUDA/TRT): 1.21.0-dev-20250228	onnxruntime-react-native: 1.21.0-dev.20250228	GPU: 1.21.0-rc1
DirectML: 1.21.0	WindowsAI: 1.21.0-dev-20250301	onnxruntime-web: 1.21.0-dev.20250206
	DirectML: 1.21.0-dev-20250228

[Areopagitics]

I tried pip install onnxruntime-gpu[cuda,cudnn]==1.21.0.dev20250228002, but it asks for a password when installing the extras.

[tjwald]

I tried adding the GPU nuget package and it wouldn't finish the restore.

dotnet nuget add source https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ORT-Nightly/nuget/v3/index.json --name ORT-Nightly

dotenet restore

Add to my csproj:

<PackageReference Include="Microsoft.ML.OnnxRuntime.Gpu" Version="1.21.0-dev-20250228"/>

And then will not complete restore.

[tianleiwu]

I tried pip install onnxruntime-gpu[cuda,cudnn]==1.21.0.dev20250228002, but it asks for a password when installing the extras.

pip install flatbuffers numpy packaging protobuf sympy
pip install --pre --index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ORT-Nightly/pypi/simple/ onnxruntime-gpu[cuda,cudnn]==1.21.0.dev20250228002

[Djdefrag]

#20713

After almost a year and 4 “major” releases this problem is still present

[SSFRPA]

#20713
If you do not intend to address this issue, please inform us in advance so that we can seek alternative solutions.