| output_dir: ./runs/pack_3_objects_plus_base/2026-06-08_23-08-07 |
| batch_size: 8 |
| num_workers: 2 |
| lr_scheduler_type: cosine |
| learning_rate: 0.0001 |
| num_epochs: 100 |
| max_steps: 20000 |
| log_every: 10 |
| save_every: 2000 |
| eval_every: 500 |
| eval_num_inference_steps: 10 |
| gradient_accumulation_steps: 1 |
| mixed_precision: bf16 |
| seed: 42 |
| max_grad_norm: 1.0 |
| weight_decay: 0.01 |
| resume: null |
| partial_resume: null |
| wandb: |
| enabled: true |
| workspace: null |
| project: fast-wam-real |
| name: pack_3_objects_plus_base |
| group: null |
| mode: online |
| data: |
| train: |
| _target_: fastwam.datasets.lerobot.robot_video_dataset.RobotVideoDataset |
| dataset_dirs: |
| - ./data/pack_3_objects_plus/perfect_lerobot |
| shape_meta: |
| images: |
| - key: cam_high |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_left_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_right_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| action: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| state: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| num_frames: 33 |
| global_sample_stride: 1 |
| action_video_freq_ratio: 4 |
| video_size: |
| - 384 |
| - 320 |
| camera_key: null |
| val_set_proportion: 0.01 |
| is_training_set: true |
| pretrained_norm_stats: null |
| skip_padding_as_possible: false |
| concat_multi_camera: robotwin |
| processor: |
| _target_: fastwam.datasets.lerobot.processors.fastwam_processor.FastWAMProcessor |
| shape_meta: |
| images: |
| - key: cam_high |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_left_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_right_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| action: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| state: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| num_obs_steps: 33 |
| num_output_cameras: 3 |
| action_output_dim: 14 |
| proprio_output_dim: 14 |
| action_state_transforms: null |
| use_stepwise_action_norm: false |
| norm_default_mode: z-score |
| norm_exception_mode: null |
| action_state_merger: |
| _target_: fastwam.datasets.lerobot.transforms.action_state_merger.ConcatLeftAlign |
| train_transforms: |
| - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor |
| - _target_: torchvision.transforms.Resize |
| size: |
| - 240 |
| - 320 |
| val_transforms: |
| - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor |
| - _target_: torchvision.transforms.Resize |
| size: |
| - 240 |
| - 320 |
| image_augmentation: true |
| train_aug_transforms: |
| cam_high: |
| - _target_: torchvision.transforms.RandomCrop |
| size: |
| - 228 |
| - 304 |
| - _target_: torchvision.transforms.Resize |
| size: |
| - 240 |
| - 320 |
| - _target_: torchvision.transforms.RandomRotation |
| degrees: 5 |
| - _target_: torchvision.transforms.ColorJitter |
| brightness: 0.3 |
| contrast: 0.4 |
| saturation: 0.5 |
| cam_left_wrist: |
| - _target_: torchvision.transforms.ColorJitter |
| brightness: 0.3 |
| contrast: 0.4 |
| saturation: 0.5 |
| cam_right_wrist: |
| - _target_: torchvision.transforms.ColorJitter |
| brightness: 0.3 |
| contrast: 0.4 |
| saturation: 0.5 |
| text_embedding_cache_dir: ./data/text_embeds_cache/pack_3_objects_plus |
| context_len: 128 |
| val: |
| _target_: fastwam.datasets.lerobot.robot_video_dataset.RobotVideoDataset |
| dataset_dirs: |
| - ./data/pack_3_objects_plus/perfect_lerobot |
| shape_meta: |
| images: |
| - key: cam_high |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_left_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_right_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| action: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| state: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| num_frames: 33 |
| global_sample_stride: 1 |
| action_video_freq_ratio: 4 |
| video_size: |
| - 384 |
| - 320 |
| camera_key: null |
| val_set_proportion: 0.01 |
| is_training_set: false |
| pretrained_norm_stats: null |
| skip_padding_as_possible: false |
| concat_multi_camera: robotwin |
| processor: |
| _target_: fastwam.datasets.lerobot.processors.fastwam_processor.FastWAMProcessor |
| shape_meta: |
| images: |
| - key: cam_high |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_left_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| - key: cam_right_wrist |
| raw_shape: |
| - 3 |
| - 480 |
| - 640 |
| shape: |
| - 3 |
| - 240 |
| - 320 |
| action: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| state: |
| - key: default |
| raw_shape: 14 |
| shape: 14 |
| num_obs_steps: 33 |
| num_output_cameras: 3 |
| action_output_dim: 14 |
| proprio_output_dim: 14 |
| action_state_transforms: null |
| use_stepwise_action_norm: false |
| norm_default_mode: z-score |
| norm_exception_mode: null |
| action_state_merger: |
| _target_: fastwam.datasets.lerobot.transforms.action_state_merger.ConcatLeftAlign |
| train_transforms: |
| - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor |
| - _target_: torchvision.transforms.Resize |
| size: |
| - 240 |
| - 320 |
| val_transforms: |
| - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor |
| - _target_: torchvision.transforms.Resize |
| size: |
| - 240 |
| - 320 |
| text_embedding_cache_dir: ./data/text_embeds_cache/pack_3_objects_plus |
| context_len: 128 |
| model: |
| _target_: fastwam.runtime.create_fastwam |
| model_id: Wan-AI/Wan2.2-TI2V-5B |
| tokenizer_model_id: Wan-AI/Wan2.2-TI2V-5B |
| dit_checkpoint_path: null |
| tokenizer_max_len: 128 |
| load_text_encoder: false |
| proprio_dim: 14 |
| redirect_common_files: false |
| mot_checkpoint_mixed_attn: false |
| action_dit_pretrained_path: checkpoints/ActionDiT_linear_interp_Wan22_alphascale_1024hdim.pt |
| skip_dit_load_from_pretrain: false |
| 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: 1.0e-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: 14 |
| action_group_causal_mask_mode: group_diagonal |
| action_dit_config: |
| action_dim: 14 |
| hidden_dim: 1024 |
| ffn_dim: 4096 |
| num_heads: 24 |
| attn_head_dim: 128 |
| num_layers: 30 |
| text_dim: 4096 |
| freq_dim: 256 |
| eps: 1.0e-06 |
| use_gradient_checkpointing: 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 |
| loss: |
| lambda_action: 1.0 |
|
|
