fix: QuantizationInfo equality operator overload by treating implicit offset as zero

[morgolock]

Update QuantizationInfo::operator== to compare normalized uniform() values when both operands are effectively per-layer, while preserving strict vector equality for per-channel quantization info.

This makes equivalent qinfo such as QuantizationInfo(1.0f) and QuantizationInfo(1.0f, 0) compare equal, avoiding incorrect dispatch differences for paths that rely on QuantizationInfo equality.

Add UNIT coverage for equivalent uniform qinfo, differing uniform qinfo, and per-channel equality behavior, and remove the pooling-specific regression coverage.

Change-Id: I54ab2cc9b2fae810b9d2767676097858eb8621c8

[gunes-arm]

Please link tickets if any

[gunes-arm]

fix: treat equivalent uniform QuantizationInfo as equal

Please start with capital letter, i.e. "fix: Treat ..."

Also, please change the PR title and description accordingly.

[gunes-arm]

fix: treat equivalent uniform QuantizationInfo as equal

Please start with capital letter, i.e. "fix: Treat ..."

Also, please change the PR title and description accordingly.

Also, what do you think about the following commit message. It rings more bells when read in release notes:?

fix: QuantizationInfo equality operator overload by treating implicit offset as zero

[morgolock]

fix: treat equivalent uniform QuantizationInfo as equal
Please start with capital letter, i.e. "fix: Treat ..."
Also, please change the PR title and description accordingly.

Also, what do you think about? It rings more bells when read in release notes:?

fix: QuantizationInfo equality operator overload by treating implicit offset as zero

Yes, sorry, I had made the mental link between the request and the release notes but somehow I forgot it again :)

[gunes-arm]

fix: treat equivalent uniform QuantizationInfo as equal
Please start with capital letter, i.e. "fix: Treat ..."
Also, please change the PR title and description accordingly.

Also, what do you think about the following commit message. It rings more bells when read in release notes:?

fix: QuantizationInfo equality operator overload by treating implicit offset as zero

What's your opinion on the commit message title? (+ PR description ought to change).
Do you have tickets that you can refer to in the description?

[morgolock]

fix: treat equivalent uniform QuantizationInfo as equal
Please start with capital letter, i.e. "fix: Treat ..."
Also, please change the PR title and description accordingly.

Also, what do you think about the following commit message. It rings more bells when read in release notes:?
fix: QuantizationInfo equality operator overload by treating implicit offset as zero

What's your opinion on the commit message title? (+ PR description ought to change). Do you have tickets that you can refer to in the description?

Apologies, sorted now. I don't have a ticket for this.

[gunes-arm]

fix: treat equivalent uniform QuantizationInfo as equal
Please start with capital letter, i.e. "fix: Treat ..."
Also, please change the PR title and description accordingly.

Also, what do you think about the following commit message. It rings more bells when read in release notes:?
fix: QuantizationInfo equality operator overload by treating implicit offset as zero

What's your opinion on the commit message title? (+ PR description ought to change). Do you have tickets that you can refer to in the description?

Apologies, sorted now. I don't have a ticket for this.

Thanks a lot Pablo. You didn't say what you think about the proposed commit message title.

[morgolock]

fix: treat equivalent uniform QuantizationInfo as equal
Please start with capital letter, i.e. "fix: Treat ..."
Also, please change the PR title and description accordingly.

Also, what do you think about the following commit message. It rings more bells when read in release notes:?
fix: QuantizationInfo equality operator overload by treating implicit offset as zero

What's your opinion on the commit message title? (+ PR description ought to change). Do you have tickets that you can refer to in the description?

Apologies, sorted now. I don't have a ticket for this.

Thanks a lot Pablo. You didn't say what you think about the proposed commit message title.

Changed the title now to fix: QuantizationInfo equality operator overload by treating implicit offset as zero for both: commit and PR

Thanks.