[v1.x] ONNX Change some 10D shapes to 8D

[Zha0q1]

ONNX Transpose requires static permutations. So to do "dynamic transpose" we sometimes need to do a trick to reshape the input tensor to a known high dimension first, do transpose, and reshape back to lower dimension.

Previously I hardcoded the high dimension to be 10-D but some inference frameworks (including trt) only supports up to 8D. This pr changes those 10D shapes to 8D

[mxnet-bot]

Hey @Zha0q1 , Thanks for submitting the PR
All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following commands:

• To trigger all jobs: @mxnet-bot run ci [all]
• To trigger specific jobs: @mxnet-bot run ci [job1, job2]

---

CI supported jobs: [unix-gpu, centos-cpu, unix-cpu, sanity, website, miscellaneous, clang, edge, windows-cpu, centos-gpu, windows-gpu]

---

Note:
Only following 3 categories can trigger CI :PR Author, MXNet Committer, Jenkins Admin.
All CI tests must pass before the PR can be merged.

[waytrue17]

LGTM, thanks for the fix!