{ "type": "WSA_Large", "n_obs_steps": 1, "input_features": { "observation.state": { "type": "STATE", "shape": [ 24 ] } }, "output_features": { "action": { "type": "ACTION", "shape": [ 24 ] } }, "device": "cuda:0", "use_amp": false, "push_to_hub": false, "repo_id": "zaleni/WSA-Large", "private": false, "tags": null, "license": null, "pretrained_path": null, "variant": "wsa_large", "model_id": "Wan-AI/Wan2.2-TI2V-5B", "tokenizer_model_id": "Wan-AI/Wan2.1-T2V-1.3B", "tokenizer_max_len": 128, "load_text_encoder": false, "redirect_common_files": true, "mot_checkpoint_mixed_attn": false, "skip_dit_load_from_pretrain": false, "action_dit_pretrained_path": "checkpoints/wsa_large/ActionDiT_linear_interp_Wan22_alphascale_1024hdim.pt", "future_3d_pretrained_path": "checkpoints/wsa_large/Future3DExpert_linear_interp_Wan22_alphascale_768hdim.pt", "native_checkpoint_path": null, "dtype": "bfloat16", "action_dim": 24, "proprio_dim": 24, "action_horizon": 32, "n_action_steps": 24, "num_inference_steps": 10, "video_dit_config": { "has_image_input": false, "patch_size": [ 1, 2, 2 ], "in_dim": 48, "hidden_dim": 3072, "ffn_dim": 14336, "freq_dim": 256, "text_dim": 4096, "out_dim": 48, "num_heads": 24, "attn_head_dim": 128, "num_layers": 30, "eps": 1e-06, "seperated_timestep": true, "require_clip_embedding": false, "require_vae_embedding": false, "fuse_vae_embedding_in_latents": true, "use_gradient_checkpointing": false, "video_attention_mask_mode": "first_frame_causal", "action_conditioned": false, "action_dim": 24, "action_group_causal_mask_mode": "group_diagonal" }, "action_dit_config": { "action_dim": 24, "hidden_dim": 1024, "ffn_dim": 4096, "num_heads": 24, "attn_head_dim": 128, "num_layers": 30, "text_dim": 4096, "freq_dim": 256, "eps": 1e-06, "use_gradient_checkpointing": false }, "future_3d_config": { "hidden_dim": 768, "ffn_dim": 3072, "text_dim": 4096, "freq_dim": 256, "eps": 1e-06, "num_heads": 24, "attn_head_dim": 128, "num_layers": 30, "num_query_tokens": 432, "da3_num_views": 3, "da3_tokens_per_view": 1296, "da3_query_dim": 2048, "query_layer_indices": [ 14, 19, 24, 29 ], "future_query_init_std": 0.02, "query_mode": "slot_noise", "query_noise_scale": 0.5, "query_noise_min_sigma": 0.0, "query_noise_max_sigma": 0.5, "query_sigma_source": "constant", "slot_pos_scale": 0.5, "use_gradient_checkpointing": false }, "lambda_3d": 0.1, "da3_variant": "large", "da3_teacher_layers": [ 11, 15, 19, 23 ], "da3_query_dim": 2048, "da3_tokens_per_view": 1296, "da3_num_views": 3, "future_3d_tokens_per_view": 144, "future_3d_view_attention_layout": "robotwin", "da3_model_path_or_name": "depth-anything/DA3-LARGE-1.1", "da3_model_name": null, "da3_code_root": null, "da3_teacher_process_res": 504, "da3_layer_weights": [ 1.0, 1.2, 1.4, 1.6 ], "future_query_init_std": 0.02, "future_3d_query_mode": "slot_noise", "future_3d_query_noise_scale": 0.5, "future_3d_query_noise_min_sigma": 0.0, "future_3d_query_noise_max_sigma": 0.5, "future_3d_query_sigma_source": "constant", "future_3d_slot_pos_scale": 0.5, "future_3d_target_index": -1, "log_da3_teacher_timing": false, "video_scheduler": { "train_shift": 5.0, "infer_shift": 5.0, "num_train_timesteps": 1000 }, "action_scheduler": { "train_shift": 5.0, "infer_shift": 5.0, "num_train_timesteps": 1000 }, "lambda_video": 1.0, "lambda_action": 1.0, "action_norm_use_stepwise": false, "action_norm_default_mode": "z-score", "action_norm_exception_mode": null, "action_stats_path": null, "optimizer_lr": 0.0001, "optimizer_betas": [ 0.9, 0.95 ], "optimizer_eps": 1e-08, "optimizer_weight_decay": 0.01, "optimizer_grad_clip_norm": 1.0, "scheduler_type": "cosine", "scheduler_warmup_steps": 5000, "scheduler_warmup_ratio": 0.05, "scheduler_decay_lr": 1e-05, "train_num_epochs": null, "train_max_steps": 300000 }