Add Test Case Demo

[yunqianevergarden]

Dear maintainers,

First of all, I would like to express my sincere gratitude for your excellent work and the elegant design of this repository. Your contributions are highly appreciated.

I have been using OfflineRL-Lib to reproduce XQL，and I find it very useful.

I would like to suggest adding a test case that demonstrates how to execute it from the command line with specific parameter settings.

And here is mine:

python run_xql_d4rl.py
--name consistent_d4rl
--task walker2d-medium-replay-v2
--seed 2
--device cuda:0
--hidden_dims [256,256]
--dropout 0.0
--conditioned_logstd False
--policy_logstd_min -5.0
--norm_layer True
--value_dropout 0.0
--num_v_update 1
--scale_random_sample 0.0
--loss_temperature 2.0
--aw_temperature 3.0
--use_log_loss False
--noise_std 0.0
--tau 0.005
--discount 0.99
--max_action 1.0
--max_clip 7.0
--actor_lr 3e-4
--critic_v_lr 3e-4
--critic_q_lr 3e-4
--actor_opt_decay_schedule cosine
--max_epoch 1000
--step_per_epoch 1000
--batch_size 256
--normalize_obs False
--normalize_reward True
--eval_interval 10
--eval_episode 10
--log_interval 10
--save_interval 50
--wandb '{"project": your project,"entity":your_wandb_entity}'
--debug False

[typoverflow]

Hi @yunqianevergarden,

Thanks a lot for the suggestion. We will add explanations on how to run the experiments from the CLI in the README.