-
Notifications
You must be signed in to change notification settings - Fork 0
/
my_PFN.py
142 lines (124 loc) · 4.69 KB
/
my_PFN.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
import torch
import torch.nn as nn
import numpy as np
class ParticleFlowNetwork(nn.Module):
r"""Parameters
----------
input_dims : int
Input feature dimensions.
num_classes : int
Number of output classes.
layer_params : list
List of the feature size for each layer.
"""
def __init__(self, input_dims, num_classes,
Phi_sizes=(100, 100, 128),
F_sizes=(100, 100, 100),
use_bn=True,
for_inference=False,
mask=True,
mask_val=0,
**kwargs):
super(ParticleFlowNetwork, self).__init__(**kwargs)
self.latent_dim=Phi_sizes[-1]
self.input_dims=input_dims
self.mask=mask
self.mask_val=mask_val
# input bn
self.input_bn = nn.BatchNorm1d(input_dims) if use_bn else nn.Identity()
# per-particle functions
phi_layers = []
for i in range(len(Phi_sizes)):
phi_layers.append(nn.Sequential(
nn.Conv1d(input_dims if i == 0 else Phi_sizes[i - 1], Phi_sizes[i], kernel_size=1),
nn.BatchNorm1d(Phi_sizes[i]) if use_bn else nn.Identity(),
nn.ReLU())
)
self.phi = nn.Sequential(*phi_layers)
# global functions
f_layers = []
for i in range(len(F_sizes)):
f_layers.append(nn.Sequential(
nn.Linear(Phi_sizes[-1] if i == 0 else F_sizes[i - 1], F_sizes[i]),
nn.ReLU())
)
f_layers.append(nn.Linear(F_sizes[-1], num_classes))
if for_inference:
f_layers.append(nn.Softmax(dim=1))
self.fc = nn.Sequential(*f_layers)
def forward(self, x):
#device=x.device
# x: the feature vector initally read from the data structure, in dimension (N, C, P)
if self.mask:
mask_matrix=(x.abs().sum(dim=1, keepdim=True) != self.mask_val)
x = self.input_bn(x)
x = self.phi(x)
if self.mask:
#x.data*=mask_matrix
#x*=mask_matrix
x=torch.mul(x,mask_matrix)
x = x.sum(-1)
return self.fc(x)
""" def forward(self, x):
device=x.device
# x: the feature vector initally read from the data structure, in dimension (N, C, P)
if self.mask:
mask_matrix=self.get_mask_matrix(x).to(device=device)
x = self.input_bn(x)
x = self.phi(x)
if self.mask:
x=x.masked_fill(mask_matrix,0)
x = x.sum(-1)
return self.fc(x)
def get_mask_matrix(self,x):
mask=(x==self.mask_val).all(dim=1,keepdim=True)
return mask """
""" def forward(self, x):
device=x.device
# x: the feature vector initally read from the data structure, in dimension (N, C, P)
if self.mask:
mask_matrix=self.get_mask_matrix(x).to(device=device)
x = self.input_bn(x)
x = self.phi(x)
if self.mask:
x=x*mask_matrix
x = x.sum(-1)
return self.fc(x)
def get_mask_matrix(self,x):
mask_2d=torch.logical_not((x==self.mask_val).all(dim=1,keepdim=True))
mask_3d=mask_2d.repeat(1,self.latent_dim,1).float()
return mask_3d """
""" def get_mask_matrix(self,x):
num_events=x.shape[0]
num_particles=x.shape[2]
judge=np.ones_like(self.input_dims)*self.mask_val
mask_=[]
if x.device.type!='cpu':
x_np=x.cpu().detach().numpy()
else:
x_np=x.detach().numpy()
x_np=x_np.reshape(num_events,num_particles,-1)
for i in range(num_events):
for j in range(num_particles):
if (x_np[i,j,:]==judge).all():
mask_.append(0)
else:
mask_.append(1)
mask_=np.array(mask_).reshape(num_events,num_particles)
return mask_ """
def get_model(data_config, **kwargs):
Phi_sizes = (128, 128, 128)
F_sizes = (128, 128, 128)
input_dims = len(data_config.input_dicts['pf_features'])
num_classes = len(data_config.label_value)
model = ParticleFlowNetwork(input_dims, num_classes, Phi_sizes=Phi_sizes,
F_sizes=F_sizes, use_bn=kwargs.get('use_bn', False))
model_info = {
'input_names': list(data_config.input_names),
'input_shapes': {k: ((1,) + s[1:]) for k, s in data_config.input_shapes.items()},
'output_names': ['softmax'],
'dynamic_axes': {**{k: {0: 'N', 2: 'n_' + k.split('_')[0]} for k in data_config.input_names}, **{'softmax': {0: 'N'}}},
}
return model, model_info
def get_loss(data_config, **kwargs):
return torch.nn.CrossEntropyLoss()