Comment:
Would it acceptable to harmonize the build string with that of tensorflow and pytorch
onnxruntime (this feedstock)
{% set build_ext = "cuda" if cuda_enabled else "cpu" %}
string: py{{ CONDA_PY }}h{{ PKG_HASH }}_{{ PKG_BUILDNUM }}_{{ build_ext }}
The main features that are missing:
- The version of cuda that things are compiled with is not immediately obvious.
cuda comes at the end, whereas it arrives at the beginning in tensorflow and pytorch
Tensorflow:
https://github.com/conda-forge/tensorflow-feedstock/blob/main/recipe/meta.yaml#L136
string: cuda{{ cuda_compiler_version | replace('.', '') }}py{{ CONDA_PY }}h{{ PKG_HASH }}_{{ PKG_BUILDNUM }} # [cuda_compiler_version != "None"]
string: cpu_py{{ CONDA_PY }}h{{ PKG_HASH }}_{{ PKG_BUILDNUM }} # [cuda_compiler_version == "None"]
Pytorch
https://github.com/conda-forge/pytorch-cpu-feedstock/blob/main/recipe/meta.yaml
string: cuda{{ cuda_compiler_version | replace('.', '') }}py{{ CONDA_PY }}h{{ PKG_HASH }}_{{ PKG_BUILDNUM }} # [cuda_compiler_version != "None"]
string: cpu_{{ blas_impl }}_py{{ CONDA_PY }}h{{ PKG_HASH }}_{{ PKG_BUILDNUM }} # [cuda_compiler_version == "None"]
Comment:
Would it acceptable to harmonize the build string with that of tensorflow and pytorch
onnxruntime (this feedstock)
The main features that are missing:
cudacomes at the end, whereas it arrives at the beginning in tensorflow and pytorchTensorflow:
https://github.com/conda-forge/tensorflow-feedstock/blob/main/recipe/meta.yaml#L136
Pytorch
https://github.com/conda-forge/pytorch-cpu-feedstock/blob/main/recipe/meta.yaml