# "halfcheetah", "hopper", "walker2d" ("medium-expert", "expert", "random")
# "pen" "hammer" "door" "relocate" ("human")
# "kitchen" ("complete" "partial" "mixed")
env_name="antmaze"
dataset="umaze-diverse"
act_norm="min_max"
rtg_norm="norm_tanh"
# "v0", "v1", "v2"
ds_version="v0"
# "RCP", "IRCP", "RWR", "RBC"
algo="IRCP"
loss_type="infoNCE"
# 1e-3 - 5e-3 - 1e-4 - 5e-4
data_noise=0.
lr=1e-4
# 0.0 - 0.01 - 0.1 - 0.5 - 0.9
alpha=-1.
seed=42
# 256 - 1024 - 2048 - 16384
batch_size=512
num_mcmc_chains=8
num_eval_episodes=10
max_iters=100
num_steps_per_iter=1000
reward_target=0.5
spectral_norm=False
save_ckpt=False
lr_schedule=False
eta=0.
scale=1.
max_g=1.
temperature=1.
filter_=False
all_grad_penalty=True
u_net=False
use_skip=True
use_bias=True
density_penalty=0.
dims=512
num_layers=8
env_dims=16
epsilon=1e-5
ema=0.
clip=0.1
init_g_randomly=True
uniform_buffer=0.05
rtg_uniform_buffer=0.05
warmup=0
grad_penalty=1.
gradient_scaling=0.
use_layer_norm=False
wandb_project="test_antmaze"
activation="relu"
wandb_entity="abi"
wab_log=True
num_action_samples=512
medium_level_traj=0
expert_level_traj=5
weight_decay=0.
rtg=False
average_reward=True



D4RL_SUPPRESS_IMPORT_ERROR=1