/
architect.py
119 lines (101 loc) · 4.79 KB
/
architect.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
import torch
import numpy as np
import torch.nn as nn
from torch.autograd import Variable
def _concat(xs):
return torch.cat([x.view(-1) for x in xs])
class Architect(object):
def __init__(self, model, args):
self.network_momentum = args.momentum
self.network_weight_decay = args.weight_decay
self.model = model
self.optimizer = torch.optim.Adam(self.model.arch_parameters(),
lr=args.arch_learning_rate,
betas=(0.5, 0.999),
weight_decay=args.arch_weight_decay)
def _compute_unrolled_model(self, input, target, eta,
network_optimizer): #,temp):
loss = self.model._loss(input, target) #,temp)
parameters = [
p for n, p in self.model.named_parameters() if p.requires_grad
]
theta = _concat(parameters).data
try:
moment = _concat(network_optimizer.state[v]['momentum_buffer']
for v in self.model.parameters()).mul_(
self.network_momentum)
except:
moment = torch.zeros_like(theta)
#names = [n for n, p in self.model.named_parameters() if not p.requires_grad]
#print(names)
dtheta = _concat(torch.autograd.grad(
loss, parameters)).data + self.network_weight_decay * theta
unrolled_model = self._construct_model_from_theta(
theta.sub(eta, moment + dtheta)) #,temp)
return unrolled_model
def step(self, input_train, target_train, input_valid, target_valid, eta,
network_optimizer, unrolled):
self.optimizer.zero_grad()
if unrolled:
self._backward_step_unrolled(input_train, target_train,
input_valid, target_valid, eta,
network_optimizer) #,temp)
else:
self._backward_step(input_valid, target_valid) #,temp)
self.optimizer.step()
def _backward_step(self, input_valid, target_valid): #,temp):
loss = self.model._loss(input_valid, target_valid) #,temp)
loss.backward()
def _backward_step_unrolled(self, input_train, target_train, input_valid,
target_valid, eta,
network_optimizer): #,temp):
unrolled_model = self._compute_unrolled_model(
input_train, target_train, eta, network_optimizer) #,temp)
unrolled_loss = unrolled_model._loss(input_valid,
target_valid) #,temp)
unrolled_loss.backward()
parameters = [
p for n, p in unrolled_model.named_parameters() if p.requires_grad
]
dalpha = [v.grad for v in unrolled_model.arch_parameters()]
vector = [v.grad.data for v in parameters]
implicit_grads = self._hessian_vector_product(vector, input_train,
target_train) #,temp)
#print(dalpha)
#print(vector)
for g, ig in zip(dalpha, implicit_grads):
g.data.sub_(eta, ig.data)
for v, g in zip(self.model.arch_parameters(), dalpha):
if v.grad is None:
v.grad = Variable(g.data)
else:
v.grad.data.copy_(g.data)
def _construct_model_from_theta(self, theta): #,temp):
model_new = self.model.new()
model_dict = self.model.state_dict()
params, offset = {}, 0
for k, v in self.model.named_parameters():
if v.requires_grad:
v_length = np.prod(v.size())
params[k] = theta[offset:offset + v_length].view(v.size())
offset += v_length
assert offset == len(theta)
model_dict.update(params)
model_new.load_state_dict(model_dict)
return model_new
def _hessian_vector_product(self, vector, input, target, r=1e-2):
R = r / _concat(vector).norm()
parameters = [
p for n, p in self.model.named_parameters() if p.requires_grad
]
for p, v in zip(parameters, vector):
p.data.add_(R, v)
loss = self.model._loss(input, target) #),temp)
grads_p = torch.autograd.grad(loss, self.model.arch_parameters())
for p, v in zip(parameters, vector):
p.data.sub_(2 * R, v)
loss = self.model._loss(input, target) #,temp)
grads_n = torch.autograd.grad(loss, self.model.arch_parameters())
for p, v in zip(parameters, vector):
p.data.add_(R, v)
return [(x - y).div_(2 * R) for x, y in zip(grads_p, grads_n)]