{ "apply_layernorm": true, "architectures": [ "EmbodiedMAEForMaskedImageModeling" ], "attention_probs_dropout_prob": 0.0, "auto_map": { "AutoConfig": "configuration_embodiedmae.EmbodiedMAEConfig", "AutoModelForMaskedImageModeling": "modeling_embodiedmae.EmbodiedMAEForMaskedImageModeling" }, "decoder_hidden_size": 512, "decoder_num_attention_heads": 8, "decoder_num_hidden_layers": 4, "dirichlet_alpha": 1.0, "drop_path_rate": 0.0, "hidden_act": "gelu", "hidden_dropout_prob": 0.0, "hidden_size": 768, "image_size": 224, "initializer_range": 0.02, "layer_norm_eps": 1e-06, "layerscale_value": 1.0, "mlp_ratio": 4, "model_type": "EmbodiedMAE", "norm_pix_loss": false, "num_attention_heads": 12, "num_hidden_layers": 12, "num_pc_centers": 196, "num_pc_knn": 64, "patch_size": 16, "qkv_bias": true, "torch_dtype": "float32", "transformers_version": "4.48.0", "unmask_sz": 98, "use_swiglu_ffn": false }