How to set the parameter to customize the number of hidden layers and their hidden sizes

[porrk]

❓ Question

Setting the config parameter seems to only make all the hidden layers the same size. How can I set the parameter to customize the number of hidden layers and their respective sizes?
For examples, if i want to set up a policy network with three hidden layers, each of size 128, 64, 32. How do I set or modify the config parameter.
I would appreciate it if you could give me a detailed reply.

Checklist

• I have checked that there is no similar issues in the repo
• I have read the documentation

[huangshiyu13]

You can pass your predefined neural network to the OpenRL:

Code Examples1:

openrl/examples/sb3/train_ppo.py

Line 37 in 8797b09

model_dict = {"model": PolicyValueNetwork}

Code Example2:

openrl/examples/nlp/train_ppo.py

Line 25 in 8797b09

model_dict = {"model": PolicyValueNetwork}

Document: https://openrl-docs.readthedocs.io/en/latest/quick_start/train_nlp.html

[porrk]

Thank you for your careful response!