Our (or to be more specific my) problem is the lack of understanding mostly everything .. we tried to find patterns / explain behavior by training different training sets (we're doing around 100-500 images at the moment) .. and we're kinda lost at this point ..
My goals are:
-) Some questions answered for the sole purpose to understand what to do and why to do and how to do.
-) Debugging a model to see why it detects certain objects
-) Retraining ... I don't know how to add classes to an already trained model
-) Train "negatives" (to battle false positives)
-) Knowledge on how to build trainingsets, how to annotate what and why
-) Validation / Testing / etc ..
-) How to pre-process images for maximum training / detection
Main goal is to get the current project as good as possible .. but also to be able to get all the projects done I can imagine in the future.
My current model is built with keras (tf-backend) and resnet50 pretrained on imagenet.
Less than 30 hrs/week
Less than 1 month< 1 monthProject LengthDuration
I am willing to pay higher rates for the most experienced freelancers