MolecularDiffusion.modules.models.shepherd_arch.model¶

Classes¶

Functions¶

`display_dict`(d[, indent])
`remap_values`(remapping_tuple, input_tensor)	# credit to: https://discuss.pytorch.org/t/cv2-remap-in-pytorch/99354/8

Module Contents¶

class MolecularDiffusion.modules.models.shepherd_arch.model.Model(params)¶

Bases: torch.nn.Module

forward(input_dict)¶

forward_decoder_joint_heterogeneous_graph_encoder(input_dict, output_dict)¶

forward_decoder_joint_processing(x_str, input_dict, output_dict)¶

forward_x1_decoder_denoiser(input_dict, output_dict)¶

forward_x1_decoder_encoder(input_dict, output_dict)¶

forward_x2_decoder_denoiser(input_dict, output_dict)¶

forward_x2_decoder_encoder(input_dict, output_dict)¶

forward_x3_decoder_denoiser(input_dict, output_dict)¶

forward_x3_decoder_encoder(input_dict, output_dict)¶

forward_x4_decoder_denoiser(input_dict, output_dict)¶

forward_x4_decoder_encoder(input_dict, output_dict)¶

decoder_joint_heterogeneous_graph_encoder = None¶

device = 'cpu'¶

exclude_variables_from_decoder_heterogeneous_graph¶

explicit_diffusion_variables¶

joint_SO3_grid¶

params¶

x1_bond_diffusion¶

MolecularDiffusion.modules.models.shepherd_arch.model.display_dict(d, indent='')¶

MolecularDiffusion.modules.models.shepherd_arch.model.remap_values(remapping_tuple, input_tensor)¶

# credit to: https://discuss.pytorch.org/t/cv2-remap-in-pytorch/99354/8

Maps integer values in input_tensor to new integer values specified by the map remapping_tuple[0]:remapping_tuple[1]

Parameters:

remapping_tuple (Tuple(torch.LongTensor, torch.LongTensor))
input_tensor (torch.LongTensor)

Returns:

(torch.LongTensor) with new values