chain_demo.py result error? about pips HOT 4 CLOSED

This is a great test, and I think you did it correctly. Some points are harder to track than others, and what you are seeing here is simply success/failure of the current model.

from pips.

thanks for reply, how to optimize further if i want apply the model in my situation?

from pips.

The simplest thing is: if you have access to some synthetic data that looks like your target data, finetune the model on that data. For example, if you want the model to excel at tracking dogs, you can train it on dogs. Right now it's more of a general-purpose machine, and making it 100% reliable is an open research problem.

from pips.

The simplest thing is: if you have access to some synthetic data that looks like your target data, finetune the model on that data. For example, if you want the model to excel at tracking dogs, you can train it on dogs. Right now it's more of a general-purpose machine, and making it 100% reliable is an open research problem.
yeah,100% reliable is an open research problem,google-deepmind/tapnet#7 (comment) should also can not agree with any more,thanks all u guys

from pips.