Functions — github.com/InhwanBae/GPGraph

CodeHub MCPcopy Create free account

hub / github.com/InhwanBae/GPGraph / functions

Functions56 in github.com/InhwanBae/GPGraph

⨍ Functions 56 ◇ Types & classes 14

↓ 8 callersMethod__init__

model_baseline.py:84

↓ 3 callersFunctiongenerate_identity_matrix

model_groupwrapper.py:171

↓ 3 callersMethodrandn

(self, n, k, d)

↓ 2 callersMethod__init__

(self, baseline_model, in_channels=2, out_channels=5, obs_seq_len=8, pred_seq_len=12, d_type=

model_groupwrapper.py:179

↓ 2 callersFunctiongraph_loss

(V_pred, V_target)

↓ 2 callersFunctionseq_to_graph

(seq_, seq_rel, pos_enc=False)

↓ 2 callersMethodsplit_heads

model_baseline.py:108

↓ 1 callersMethod__init__

(self, stack_n=1000, fast_sample=True)

↓ 1 callersFunctionbivariate_loss

(V_pred,V_trgt)

↓ 1 callersFunctioncompute_batch_metric

Get ADE, FDE, TCC scores for each pedestrian

↓ 1 callersMethodfind_group_indices

(self, v, dist_mat)

model_groupwrapper.py:21

↓ 1 callersFunctiongenerate_statistics_matrices

r"""generate mean and covariance matrices from the network output.

↓ 1 callersMethodgroup_backprop_trick_threshold

The main trick for hard is to do (v_hard - v_soft).detach() + v_soft Sample hard categorical using "Straight-through" trick I

model_groupwrapper.py:52

↓ 1 callersFunctionloc_pos

↓ 1 callersFunctionmain

↓ 1 callersMethodped_group_mask

model_groupwrapper.py:111

↓ 1 callersMethodped_group_pool

model_groupwrapper.py:94

↓ 1 callersMethodped_group_unpool

model_groupwrapper.py:106

↓ 1 callersFunctionpoly_fit

Input: - traj: Numpy array of shape (2, traj_len) - traj_len: Len of trajectory - threshold: Minimum error to be considered for non l

↓ 1 callersFunctionread_file

(_path, delim='\t')

↓ 1 callersFunctiontest

(SAMPLES=20, TRIALS=100)

↓ 1 callersFunctiontrain

(epoch, model, optimizer, checkpoint_dir, loader_train)

↓ 1 callersFunctionvalid

(epoch, model, checkpoint_dir, loader_val)

Method__getitem__

Args: - data_dir: Directory containing dataset files in the format <frame_id> <ped_id> <x> <y> - obs_len: Number of t

(self, d_type='learned', th=1., in_channels=16, hid_channels=32, n_head=1, dropout=0)

model_groupwrapper.py:6

(self, mix_type='mean', n_mix=3, out_channels=5, pred_seq_len=12)

model_groupwrapper.py:121

(self, in_cha, out_cha)

model_baseline.py:11

(self, number_asymmetric_conv_layer=7, spatial_channels=4, temporal_channels=4)

model_baseline.py:39

(self, in_dims=2, d_model=64, num_heads=4)

model_baseline.py:96

(self, obs_len=8)

model_baseline.py:145

(self, spa_in_dims=2, tem_in_dims=3, embedding_dims=64, obs_len=8, dropout=0, number_asymmetr

model_baseline.py:163

(self, in_dims=2, embedding_dims=16, dropout=0)

model_baseline.py:217

(self, in_dims=16, embedding_dims=16, dropout=0)

model_baseline.py:237

(self, number_asymmetric_conv_layer=7, embedding_dims=64, number_gcn_layers=1, dropout=0,

model_baseline.py:277

(predAll,targetAll,count_)

Functioncloser_to_zero

(current,new_v)

(predAll,targetAll,count_)

Methodfind_group_indices_ratio

(self, v, dist_mat)

model_groupwrapper.py:36

(self, v, v_abs, tau=0.1, hard=True)

model_groupwrapper.py:64

(self, v_stack)

model_groupwrapper.py:134

(self, v_abs, v_rel)

model_groupwrapper.py:198

model_baseline.py:26

(self, dense_spatial_interaction, dense_temporal_interaction, threshold=0.5)

model_baseline.py:58

(self, x, dim=0, eps=1e-5)

model_baseline.py:87

(self, x, mask=False, multi_head=False)

model_baseline.py:116

model_baseline.py:155

(self, graph, identity, mask=None)

model_baseline.py:182

(self, graph, adjacency)

model_baseline.py:225

(self, graph, normalized_spatial_adjacency_matrix, normalized_temporal_adjacency_matrix)

model_baseline.py:251

(self, graph, identity, mask=None)

model_baseline.py:313

Functiongenerate_adjacency_matrix

model_groupwrapper.py:152

Functionnodes_rel_to_nodes_abs

(nodes,init_node)

Functionseq_to_nodes

(seq_,max_nodes = 88)