Why are the MLP parameters here so large?

[SanMumumu]

Hi,

The role of MLP as I understand is used to decode the representation of encoder3d, function similar to decoder.

If the mlp parameter is too large, will it affect the learning of encoder3d and the robotic arm?

GNFactor/GNFactor/agents/gnfactor_bc/models_embed.py

Lines 359 to 378 in bc9cbdd

	if coarse or self.mlp_fine is None:
	mlp_output, prev_output = self.mlp_coarse(
	mlp_input,
	combine_inner_dims=(self.num_views_per_obj, B),
	combine_index=combine_index,
	dim_size=dim_size,
	ret_last_feat=ret_last_feat,
	language_embed=self.language,
	batch_size=bs
	)
	else:
	mlp_output, prev_output = self.mlp_fine(
	mlp_input,
	combine_inner_dims=(self.num_views_per_obj, B),
	combine_index=combine_index,
	dim_size=dim_size,
	ret_last_feat=ret_last_feat,
	language_embed=self.language,
	batch_size=bs
	)

GNFactor/GNFactor/conf/method/GNFACTOR_BC.yaml

Lines 129 to 135 in bc9cbdd

	mlp:
	n_blocks: 5
	d_hidden: 512
	combine_layer: 3
	combine_type: average
	beta: 0.0
	use_spade: False