-
Notifications
You must be signed in to change notification settings - Fork 115
How are the model cpp file generated? #26
Comments
also, could i possibly train myself with my own card images? |
@ericlewis those are entirely appropriate questions. Unfortunately, the people most qualified to answer them (mainly @josharian, plus myself as junior lieutenant) have so far failed to accomplish our intended goal of cleaning up and posting the many complicated Python scripts involved, plus documenting them well enough so that they would be of any use to other developers. Rather more unfortunately, neither of us is currently actively involved with card.io. For an overview of part of what was involved, see expiry_categorization.md. |
@ericlewis, did you ever have any luck with this? I was thinking about attempting some local training and was wondering how many training images one would need. |
Sadly, no. We ended up moving on to a different style of detection for what we needed.
…Sent from my iPhone
On Mar 14, 2017, at 1:29 PM, Ace Ellett ***@***.***> wrote:
@ericlewis, did you ever have any luck with this? I was thinking about attempting some local training and was wondering how many training images one would need.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
|
The conv models, how are they generated?
The text was updated successfully, but these errors were encountered: