areusch commented on issue #9453: URL: https://github.com/apache/tvm/issues/9453#issuecomment-964763009
@baibeta could you post the `microtvm_api_server.py` you're using? in general this is intentional: - it's expected that the user will supply a template project for use with microTVM - TVM will always try to invoke generate_project with that template project, passing the MLF, and this function is expected to create a new project. it's expected that `microtvm_api_server.py` will copy itself into the new project. - TVM will then close the connection to the `microtvm_api_server.py` and open a connection to the newly-copied `microtvm_api_server.py` in the generated project. - it's expected that this new copy has a way to detect that it was copied (e.g. you can also set a special file in the generated project, place the MLF tar in the generated project, or look for sources in the generated project which came from the MLF). - the new copy should then return is_template=false. let me know if this doesn't make sense. please feel free to re-open if there's an error here, happy to address any concern you may have. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
