#DQN Agent Playing minetester-treechop_shaped-v0

This is a trained model of a DQN agent playing minetester-treechop_shaped-v0. The model was trained by using minetest-baselines.

Command to reproduce the training

python -m minetest_baselines.jax.train_dqn_cleanrl.py --exp-name dqn_treechop-v0_fixedseed --seed 1 --capture-video --track --wandb-entity rkla --learning-rate 3e-4 --save-model --upload-model --hf-entity rkla --buffer-size 100000 --exploration-fraction 0.8 --learning-rate 0.0003 --total-timesteps 1000000

Hyperparameters

{'batch_size': 128,
 'buffer_size': 100000,
 'capture_video': True,
 'end_e': 0.01,
 'env_id': 'minetester-treechop_shaped-v0',
 'exp_name': 'dqn_treechop-v0_fixedseed',
 'exploration_fraction': 0.8,
 'gamma': 0.99,
 'hf_entity': 'rkla',
 'learning_rate': 0.0003,
 'learning_starts': 5000,
 'num_envs': 1,
 'save_model': True,
 'seed': 1,
 'start_e': 1,
 'target_network_frequency': 10000,
 'tau': 1.0,
 'total_timesteps': 1000000,
 'track': True,
 'train_frequency': 10,
 'upload_model': True,
 'wandb_entity': 'rkla',
 'wandb_project_name': 'cleanRL'}