behaviors:
My Behavior:
trainer_type: ppo
hyperparameters:
batch_size: 128
buffer_size: 2048
learning_rate: 0.0005
beta: 0.01
epsilon: 0.2
lambd: 0.95
num_epoch: 5
learning_rate_schedule: linear
network_settings:
normalize: true
hidden_units: 128
num_layers: 2
vis_encode_type: nature_cnn
- Batch size: The batch size has been decreased to 128 to improve stability during training.
- Learning rate: The learning rate has been increased to 0.0005 to help the agent learn more quickly.
- Number of epochs: The number of training epochs per update has been increased to 5 to give the agent more opportunities to learn from its experiences.
- Network architecture: The network now uses the nature_cnn visual encoder, which is a good choice for image-based observations. The number of hidden units has been reduced to 128 to improve training stability.
- Normalization: Normalization has been enabled, which can help the agent learn more effectively by reducing the impact of different scales and offsets in the observations.
