search

deeplearninc / relaax

Reinforcement Learning framework to facilitate development and use of scalable RL algorithms and applications
Other
62 stars 10 forks source link

Serve the model #99

Closed tsaoyu closed 6 years ago

tsaoyu commented 6 years ago

I have created and trained my model on my own environment. My question is how to serve the trained model with relaax? Is there a CLI option or similar? Many thanks.

4SkyNet commented 6 years ago

Hi @tsaoyu!

Your model should be relaax compatible to use it further. We use standard tensorflow checkpoints to store & restore the training process (you can find respective folders checkpoints && metrics when training is initiated), since that, if you perform training with not-relaax (at least tensor compatible) model you should transfer the weight by hand with your own procedure.

p.s> sorry for some delay with reply

tsaoyu commented 6 years ago

@4SkyNet Many thanks! I management to run the model by turn off the 'exploit' feature. I will see if I could manually transfer the weights later. Cheers