Tool/software:
Dear Expert,
Yesterday, I trained an object detection model with 104 images in it. But the inference is not working well. One keyword appearing by an interval of time.
Today, to test I created a sample OD project that has around 25 images and trained in same way. It's inference is 70% around good. I thought that it can be more accurate by feeding more images.
Now I added more images in the yesterday trained model and retrying. Unfortunately Training takes more time than usual and has low performance graph.
Can you please help me to resolve this issue. Post compilation accuracy is around 50% only.
One more thing, Anything related to image format/ Quality or something else depend the model accuracy,
How can I effectively train a model. Should it need same number of images for all annotation classes.
Warm Regards,
Sajan