tztsai / Energy-Efficient-5G-RL

This repository presents a multi-agent reinforcement learning approach for energy-efficient collaborative control of base stations in 5G networks.
MIT License
27 stars 3 forks source link

DQN replaybuffer missing #9

Closed LLjiko closed 4 months ago

LLjiko commented 7 months ago

hello Bro,I really appreciate the code you provided. But I cant find the replaybuffer of DQN. The mistake comes from " from stable_baselines3.common.buffers import ReplayBuffer". It seems no such file in your project.

tztsai commented 7 months ago

That's a third party library for reinforcement learning. You need to install it using pip install stable-baselines3.

LLjiko commented 7 months ago

Ok.Thanks

nourelhouda1997 commented 7 months ago

hi sir , i don't know how to open train.sh can you provide more information for me. thanks

LLjiko commented 7 months ago

You can use the command line and input train.sh in the dictionary of the project. Or you can input python train.py --algorithm_name mappo --experiment_name check1 --scenario RANDOM --accelerate 1200 --seed 1 --n_training_threads 4 --n_rollout_threads 42 --num_mini_batch 1 --num_env_steps 1512000 --ppo_epoch 10 --gain 0.01 --gamma 0.99 --lr 5e-4 --critic_lr 5e-4 --value_loss_coef 1 --log_level NOTICE --log_interval 1 --w_qos 4 --w_xqos 0.005 in the dictionary of the project.

nourelhouda1997 commented 7 months ago

i don't understand sorry can you give more you know this my first time in this work and i'm traying to learn so please help me