metadata
language: en
library_name: torch
license: mit
tags:
- table-2
Model Card for ahalev/mcuu-table-2-1o7w33o3
This model corresponds to run(s) in Table 2, specifically that with the hyperparameters:
1) {'scenario': 6, 'forecast_horizon': 24, 'intrinsic_reward_weight': 0.0001, 'bound_reward_weight': 'cosine', 'noise_std': 0.01}
Usage
from trainer import Trainer
trainer = Trainer.from_pretrained('ahalev/mcuu-table-2-1o7w33o3')
algo, env = trainer.algo, trainer.env
# Get an action from a random observation
action, _ = algo.policy.get_action(env.observation_space.sample())
# Evaluate the policy over 2920 timesteps
evaluation = trainer.evaluate()
For more information, see the repo and the paper.
This model was created by @ahalev.