tools/update_doc.sh, the urls in
data.jsonare edited to opset 11 (https://github.com/onnx/onnx/blob/master/onnx/backend/test/cmd_tools.py#L36 and https://github.com/onnx/onnx/blob/master/onnx/backend/test/case/model/__init__.py#L61), so that the invalid urls are read in https://github.com/onnx/onnx/blob/master/onnx/backend/test/loader/__init__.py#L45
tools/update_doc.shedits the urls to opset 11 (in cmd_tools.py), and then tries to download the wrong urls (in stat_coverage.py). The download will be skipped if there are already models of the same name on the disk. This could be the reason why the problem got ignored.
Our next SIG meeting is Wednesday Sept 11 at 9pm pacific time. Here the details:
MS Team link:
+1 425-616-0754 United States, Seattle (Toll)
Conference ID: 720 240 30#
Cin Gemm optional --- onnx/onnx#2330. Does it make sense to everyone?
This is an announcement from the ONNX Foundation Working Group. We have created a poll to collect preferences from the ONNX community on what non-profit foundation we should join.
Please make your voice heard by voting here: https://forms.gle/z4zEs2MGcQiV6tar9
The options are Linux Foundation, Apache or Eclipse. Our selection and interview process is shared in the link above.
Jim Spohrer (IBM) and Ryan Loney (Intel)
Co-chairs, ONNX Foundation WG