Adding 64 bit tensor data types #54
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In testing this interface with commonly available models, there are data types missing in the specification. I have added here both integer and floating point 64-bit types as these are defined in the WASM specification as native types.
There are many other types defined by the frameworks listed in the specification. Is the goal to provide an interface which is a superset of these types or simply represent the most common. For example, many of the quantized types are primarily used internally, not at the input/output boundary.