Migrate towards autogen code #22
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Generate all MNIST operators with the Python script and start integrating them in the existing code. So far only float type is supported
-Python generated code is now being used -MNIST and tinyYOLO models are working -Opset and function resolver is not hardcoded anymore. TODOs for next patches (see issues) -Remove unused code -Remove all switch cases in the operator (no longer needed) -There is a minor issue when resolving the operator functions. When a function is not defined (i.e. for double) there is an error because that symbol can't be found. This should be linked with the default operator stub (using weakrefs?) -There is a workaround with the operator set. opset=12 is hardcoded for the initial tests.
Modify manually maxpool resolver. The Python generator for resolvers has to be rethought. Doesn't make sense for some operators. This modification is a temporal fix until we figure out how to handle this.
Traces to debug issue in win/linux
More traces
And more traces
Fix data_type for outputs
Windows CI was unexpectedly broken. Attempting to fix it.
Generate all MNIST operators with the Python script and start integrating them in the existing code. So far only float type is supported
-Python generated code is now being used -MNIST and tinyYOLO models are working -Opset and function resolver is not hardcoded anymore. TODOs for next patches (see issues) -Remove unused code -Remove all switch cases in the operator (no longer needed) -There is a minor issue when resolving the operator functions. When a function is not defined (i.e. for double) there is an error because that symbol can't be found. This should be linked with the default operator stub (using weakrefs?) -There is a workaround with the operator set. opset=12 is hardcoded for the initial tests.
Modify manually maxpool resolver. The Python generator for resolvers has to be rethought. Doesn't make sense for some operators. This modification is a temporal fix until we figure out how to handle this.
Traces to debug issue in win/linux
More traces
And more traces
Fix data_type for outputs
…evuelta/embedded-ml into migrate-towards-autogen-code
Remove unused code
Fix return
|
@nopeslide I modified a bit the Python scripts to comply with the new interface and created the functions for MNIST and YOLO. The resolving is not hardcoded anymore and the resolver functions generated by Python are being used now. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Migrate to autogenerated code
TODOs for next patches (see issues)