| { |
| "model": { |
| "obs_dim": 4, |
| "action_dim": 2, |
| "device": "cuda", |
| "archi": { |
| "z_dim": 100, |
| "norm_z": true, |
| "f": { |
| "hidden_dim": 1024, |
| "model": "simple", |
| "hidden_layers": 1, |
| "embedding_layers": 2, |
| "num_parallel": 2, |
| "ensemble_mode": "batch" |
| }, |
| "b": { |
| "hidden_dim": 256, |
| "hidden_layers": 2, |
| "norm": true |
| }, |
| "actor": { |
| "hidden_dim": 1024, |
| "model": "simple", |
| "hidden_layers": 1, |
| "embedding_layers": 2, |
| "learn_std": false, |
| "per_state_std": false, |
| "squash": false |
| }, |
| "critic": { |
| "hidden_dim": 1024, |
| "hidden_layers": 1, |
| "embedding_layers": 2, |
| "num_parallel": 2, |
| "ensemble_mode": "batch" |
| } |
| }, |
| "inference_batch_size": 500000, |
| "seq_length": 1, |
| "actor_std": 0.2, |
| "norm_obs": false |
| }, |
| "train": { |
| "lr_f": 0.0001, |
| "lr_b": 1e-06, |
| "lr_actor": 1e-06, |
| "weight_decay": 0.0, |
| "clip_grad_norm": 0.0, |
| "fb_target_tau": 0.01, |
| "ortho_coef": 1.0, |
| "train_goal_ratio": 0.5, |
| "fb_pessimism_penalty": 0.0, |
| "actor_pessimism_penalty": 0, |
| "stddev_clip": 0.3, |
| "q_loss_coef": 0.0, |
| "batch_size": 1024, |
| "discount": 0.99, |
| "update_z_every_step": 150, |
| "z_buffer_size": 10000, |
| "h_coef": 1.0 |
| }, |
| "cudagraphs": true, |
| "compile": true |
| } |
