25 made first implementation of a MMIF wrapper of the model we trained. One of the implementational choice (given time constraint at the time) was hard-coding model configuration into classify.py. I request a new feature that basically decouple the model configuration to another generic file (could be .py, .json. or .yaml) thats' imported into classify.py and in turn into app.py if necessary.
Moreover, if we can generate such a file from a result report file (such as this, note that the file contents is somewhat outdated now, and the latest train.py requires (and records) more configuration k-v)
New Feature Summary
25 made first implementation of a MMIF wrapper of the model we trained. One of the implementational choice (given time constraint at the time) was hard-coding model configuration into
classify.py
. I request a new feature that basically decouple the model configuration to another generic file (could be.py
,.json.
or.yaml
) thats' imported intoclassify.py
and in turn intoapp.py
if necessary.Moreover, if we can generate such a file from a result report file (such as this, note that the file contents is somewhat outdated now, and the latest
train.py
requires (and records) more configuration k-v)Related
No response
Alternatives
No response
Additional context
No response