Added Squeeze operator

[bezineb5]

Hi,

I've added the Squeeze operator, following a previous PR on Unsqueeze.
Specifications are here: https://github.com/onnx/onnx/blob/master/docs/Operators.md#squeeze

I'll try to add more tests, but do you have any tips to generate the model and put it in binary in the test?

Thanks!

[codecov]

Codecov Report

Merging #124 into master will increase coverage by 0.03%.
The diff coverage is 92.72%.

@@            Coverage Diff             @@
##           master     #124      +/-   ##
==========================================
+ Coverage   85.38%   85.42%   +0.03%     
==========================================
  Files         392      393       +1     
  Lines       10279    10334      +55     
==========================================
+ Hits         8777     8828      +51     
- Misses       1316     1318       +2     
- Partials      186      188       +2

Impacted Files	Coverage Δ
backend/x/gorgonnx/squeeze.go	92.72% <92.72%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 611bf02...b1c822f. Read the comment docs.

[owulveryck]

Hello,

Thanks for this PR;
Some tests are auto generated by the testbackend package in gorgonnx.
All the operators are tested and those triggering an ErrNotImplemented ate skipped.

Therefore, your implementation should have been tested by the CI already (and the CI is green so I assume the code is ok)

You can try it manually by running go test -run=ONNX/Unsqueeze -v within gorgonnx.

I will review the code and merge the PR; thanks again.

[owulveryck]

The tests of the init functions are usually not covered. We may need to add something to test them easily. WDYT?

[bezineb5]

Sure, I'll take care of that.

[owulveryck]

You can try it manually by running go test -run=ONNX/Unsqueeze -v within gorgonnx.

I mean: go test -run=ONNX/Squeeze -v

[bezineb5]

Done - I added unit tests for the various attributes which could be found in an ONNX.