I am working on a simple project on ninjatrader. I created a reinforcement learning algorithm and want to use the Ninjatrader optimizer to train it. I already have all the classes created( agent, memory, brain, trainer, and optimizer classes), but need some debugging. The algorithm saves a copy of my trainer(with agent and all parameters) after each epoch, but crashes when its time to reload the trainer class. The issue is that for each epoch, the strategy is re-started. Because of this, I need to momentarily serialize my agent, and the settings, while the strategy resets to its initial state. If I don’t serialize it, the agent starts training from scratch again, which I want to avoid). The agent is a deep network from a 3rd party library.
14 pekerja bebas membida secara purata $185 untuk pekerjaan ini
hello how are you? I read your project discretion carefully. I can do your project actively. if you hire me, I will do my best for your project. Best Regards.