transformer/config.json

{
  "patch_size": [
    1,
    2,
    2
  ],
  "num_attention_heads": 24,
  "attention_head_dim": 128,
  "in_channels": 48,
  "out_channels": 48,
  "action_dim": 20,
  "text_dim": 4096,
  "freq_dim": 256,
  "ffn_dim": 14336,
  "num_layers": 30,
  "cross_attn_norm": true,
  "eps": 1e-06,
  "rope_max_seq_len": 1024,
  "pos_embed_seq_len": null,
  "attn_mode": "flex",
  "tactile_in_channels": 3,
  "tactile_num_tokens": 4,
  "tactile_encoder_dim": 256,
  "max_tactile_streams": 4,
  "use_local_tactile": false,
  "use_proprio_state": false,
  "use_contact_gate": false,
  "contact_gate_layers": 2,
  "contact_gate_heads": 8,
  "contact_gate_stop_grad": false,
  "is_mot": true,
  "mot_expert_ffn_dim": {
    "action": 4096,
    "tactile": 4096
  },
  "mot_expert_hidden_dim": {
    "action": 1024,
    "tactile": 1024
  },
  "mot_cross_attn_experts": [
    "video",
    "action"
  ]
}