LycorisQ is a neat reinforcement learning framework based on LycorisNet.
- Universal and practical.
- Q network automation design.
- End to end.
The project is based on LycorisNet, and the installation of LycorisNet can be found here.
pip install LycorisQ
The APIs provided by Agent (from LycorisQ import Agent):
| Function | Description | Inputs | Returns |
|---|---|---|---|
| Agent(config) | Constructor. | config: The configuration information, including 11 configuration fields. | An object of the class Agent. |
| train(data) | Sampling data from the sample pool to train the neural network. | data: Each input data has 5 dimensions, which are action, reward, current_state, next_state, and done. | |
| evaluate(data) | Input observation to evaluate the corresponding strategy. | data: Observation. Environmental information. | The scores corresponding to the actions, in the form of a list. |
| save(path1, path2) | Save the model and related configurations. | path1: The path to store the model. path2: The path to store the configurations. |
|
@staticmethodload(path1, path2) |
Import pre-trained models and related configurations. | path1: The path to import the model. path2: The path to import the configurations. |
|
| set_config(config) | Set the configuration information of Agent. | config: The configuration information, including 11 configuration fields. | |
| set_lr(learning_rate) | Set the learning rate of the neural network. | learning_rate: The learning rate of the neural network. | |
| set_workers(workers) | Set the number of worker threads to train the model. | workers: The number of worker threads. | |
@staticmethodversion() |
Returns the version information of Agent. |
Configurable fields:
| Field | Description | Default |
|---|---|---|
| capacity | Capacity of LycorisNet. | |
| state_dim | Dimension of the state. | |
| action_dim | Dimension of the action. | |
| nodes | The number of hidden nodes added for each neural network. | |
| connections | The number of connections added for each neural network. | |
| depths | Total layers of each neural network. | |
| batch_size | Batch size. | |
| memory | The capacity of the sample pool. | |
| evolution | Number of LycorisNet evolutions. | 0 |
| verbose | Whether to output intermediate information. | False |
| gamma | A parameter in Q-learning. | 0.9 |
- CartPole
- MountainCar
- More examples will be released in the future.
LycorisQ is released under the LGPL-3.0 license. By using, distributing, or contributing to this project, you agree to the terms and conditions of this license.