Custom Obj detection algorithm on Sagemaker help
Budget: $750 – $1,500 AUD
TLDR: I have photos annotated, need help with getting to next step (getting model artifact that can be used in aws neo)
I have images that have been labeled with ground truth in the augmented manifest format (happy to transfer into another format such as TFRecord). I am requiring help with developing a custom algorithm to train the model that will be able to be optimised through AWS Neo. Don't need any hyperparameter optimization, only an algorithm to train model artifacts. Not too fussy about the framework used. Hoping (but not necessary) to use Jupiter notebook, possibly ONNX.
Only have around 2000 images annotated, but want a solution that could scale.
I have images that have been labeled with ground truth in the augmented manifest format (happy to transfer into another format such as TFRecord). I am requiring help with developing a custom algorithm to train the model that will be able to be optimised through AWS Neo. Don't need any hyperparameter optimization, only an algorithm to train model artifacts. Not too fussy about the framework used. Hoping (but not necessary) to use Jupiter notebook, possibly ONNX.
Only have around 2000 images annotated, but want a solution that could scale.