How to create a continuous learning ML/DL model for Computer Vision? Which approach would be ameliorated to train the ML/DL model up-to-date with respect to the new images/dataset? Let's suppose if new images could have come on a daily basis, then would I need to re-train the complete model again with the new dataset(Images)? Please let me know, Is there any way to train the model with new images & merge into the old trained model?
You may need to use Regression models for continuous learning, see for example:
https://statisticsbyjim.com/regression/choosing-regression-analysis/
https://pdfs.semanticscholar.org/149d/16329d4ebb8044c39456b6e682b257443536.pdf
Article Statistics Notes: The cost of dichotomising continuous variables
Hi Muhammad,
Thanks for sharing the articles on the same. But these links should work fine when the dataset to be available;le on Numbers. Actually I am wanting to merge a new frozen graph into an existing trained frozen graph that will optimize the cost outlay of the hardware.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images