in python/dgllife/model/pretrain/moleculenet/esol.py [0:0]
def create_esol_model(model_name):
"""Create a model.
Parameters
----------
model_name : str
Name for the model.
Returns
-------
Created model
"""
n_tasks = 1
if model_name == 'GCN_canonical_ESOL':
dropout = 0.0004181672129021179
return GCNPredictor(in_feats=74,
hidden_feats=[128],
activation=[F.relu],
residual=[True],
batchnorm=[False],
dropout=[dropout],
predictor_hidden_feats=1024,
predictor_dropout=dropout,
n_tasks=n_tasks)
elif model_name == 'GCN_attentivefp_ESOL':
dropout = 0.03400405080274294
return GCNPredictor(in_feats=39,
hidden_feats=[64],
activation=[F.relu],
residual=[False],
batchnorm=[False],
dropout=[dropout],
predictor_hidden_feats=256,
predictor_dropout=dropout,
n_tasks=n_tasks)
elif model_name == 'GAT_canonical_ESOL':
dropout = 0.28070328302954156
return GATPredictor(in_feats=74,
hidden_feats=[32],
num_heads=[4],
feat_drops=[dropout],
attn_drops=[dropout],
alphas=[0.4994779445224584],
residuals=[True],
biases=[False],
predictor_hidden_feats=16,
predictor_dropout=dropout,
n_tasks=n_tasks)
elif model_name == 'GAT_attentivefp_ESOL':
dropout = 0.00033036046538620356
return GATPredictor(in_feats=39,
hidden_feats=[32],
num_heads=[8],
feat_drops=[dropout],
attn_drops=[dropout],
alphas=[0.7197105722372982],
residuals=[False],
biases=[False],
predictor_hidden_feats=32,
predictor_dropout=dropout,
n_tasks=n_tasks)
elif model_name == 'Weave_canonical_ESOL':
return WeavePredictor(node_in_feats=74,
edge_in_feats=13,
num_gnn_layers=3,
gnn_hidden_feats=256,
graph_feats=128,
gaussian_expand=True,
n_tasks=n_tasks)
elif model_name == 'Weave_attentivefp_ESOL':
return WeavePredictor(node_in_feats=39,
edge_in_feats=11,
num_gnn_layers=1,
gnn_hidden_feats=32,
graph_feats=256,
gaussian_expand=False,
n_tasks=n_tasks)
elif model_name == 'MPNN_canonical_ESOL':
return MPNNPredictor(node_in_feats=74,
edge_in_feats=13,
node_out_feats=32,
edge_hidden_feats=64,
num_step_message_passing=3,
num_step_set2set=2,
num_layer_set2set=3,
n_tasks=n_tasks)
elif model_name == 'MPNN_attentivefp_ESOL':
return MPNNPredictor(node_in_feats=39,
edge_in_feats=11,
node_out_feats=32,
edge_hidden_feats=64,
num_step_message_passing=1,
num_step_set2set=2,
num_layer_set2set=2,
n_tasks=n_tasks)
elif model_name == 'AttentiveFP_canonical_ESOL':
return AttentiveFPPredictor(node_feat_size=74,
edge_feat_size=13,
num_layers=3,
num_timesteps=5,
graph_feat_size=16,
dropout=0.3144543143291027,
n_tasks=n_tasks)
elif model_name == 'AttentiveFP_attentivefp_ESOL':
return AttentiveFPPredictor(node_feat_size=39,
edge_feat_size=11,
num_layers=5,
num_timesteps=5,
graph_feat_size=16,
dropout=0.19597186400407615,
n_tasks=n_tasks)
elif model_name == 'gin_supervised_contextpred_ESOL':
jk = 'last'
model = GINPredictor(
num_node_emb_list=[120, 3],
num_edge_emb_list=[6, 3],
num_layers=5,
emb_dim=300,
JK=jk,
dropout=0.5,
readout='sum',
n_tasks=n_tasks
)
model.gnn.JK = jk
return model
elif model_name == 'gin_supervised_infomax_ESOL':
jk = 'sum'
model = GINPredictor(
num_node_emb_list=[120, 3],
num_edge_emb_list=[6, 3],
num_layers=5,
emb_dim=300,
JK=jk,
dropout=0.5,
readout='sum',
n_tasks=n_tasks
)
model.gnn.JK = jk
return model
elif model_name == 'gin_supervised_edgepred_ESOL':
jk = 'max'
model = GINPredictor(
num_node_emb_list=[120, 3],
num_edge_emb_list=[6, 3],
num_layers=5,
emb_dim=300,
JK=jk,
dropout=0.5,
readout='mean',
n_tasks=n_tasks
)
model.gnn.JK = jk
return model
elif model_name == 'gin_supervised_masking_ESOL':
jk = 'last'
model = GINPredictor(
num_node_emb_list=[120, 3],
num_edge_emb_list=[6, 3],
num_layers=5,
emb_dim=300,
JK=jk,
dropout=0.5,
readout='mean',
n_tasks=n_tasks
)
model.gnn.JK = jk
return model
else:
return None