test add tests for float16 #496
Conversation
| } | ||
| }, | ||
| { | ||
| "name": "add float32 2D tensors", |
There was a problem hiding this comment.
Hmm, why are all these other test cases being deleted? Don't we want at least test cases for add that still tests float32?
There was a problem hiding this comment.
It's just to clearly show test for float16.
There was a problem hiding this comment.
The PR title shows that float16 tests are being added, but I see float32 tests being deleted 🤔. Don't we want add tests to include both float16 and float32 cases?
There was a problem hiding this comment.
The PR title shows that float16 tests are being added
Actually here "add" means add operation.
Don't we want add tests to include both float16 and float32 cases?
Right, adding tests for float32，float16, and other types of https://webmachinelearning.github.io/webnn/#enumdef-mloperandtype are our target.
There was a problem hiding this comment.
My question remains - why don't we just keep some of the float32 test cases for the add operator instead of deleting the existing ones? If you delete the float32 test cases, do you plan to re-add them later in another pull request?
There was a problem hiding this comment.
It's a sample to only show a test for float16.
There was a problem hiding this comment.
Wait a minute, this is your own private repo (BruceDai/wpt), not web-platform-tests - that was the missing detail! So then I don't even need to sign off since this is your repo anyway, but yeah, it's fine to me.
@fdwr PTAL, thanks.