disentangling_categorization/optim.py at main · FICS/disentangling_categorization · GitHub

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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
import torch
import logging
log = logging.getLogger('optimizers')


def set_eval(models):
    for m in models:
        m.eval()


def set_train(models):
    for m in models:
        m.train()


def count_params(model):
    return sum([p.data.nelement() for p in model.parameters()])


def disable_parameter_requires_grad(model):
    for param in model.parameters():
        param.requires_grad = False


def enable_parameter_requires_grad(model):
    for param in model.parameters():
        param.requires_grad = True


def log_whole_system_params(sender_percept, recv_percept, _game, log=print):
    log('Optimized parameters:')
    log('sender_percept')
    for name, param in sender_percept.named_parameters():
        if param.requires_grad:
            log(f"\t{name}")

    log('sender')
    for name, param in _game.sender.named_parameters():
        if param.requires_grad:
            log(f"\t{name}")

    log('recv_percept')
    for name, param in recv_percept.named_parameters():
        if param.requires_grad:
            log(f"\t{name}")

    log('receiver')
    for name, param in _game.receiver.named_parameters():
        if param.requires_grad:
            log(f"\t{name}")


def agents_only(sender_percept, sender, receiver, log=print):
    sender_percept.model_multi.eval()
    sender.train()
    receiver.train()

    sender_percept.choose_grad('off', log=log)
    sender.choose_grad('on')
    receiver.choose_grad('on')

    log('\tSystem configuration: agents only')


def semiosis_joint(sender_percept, sender, receiver, log=print):
    sender_percept.model_multi.train()
    sender.train()
    receiver.train()

    sender_percept.choose_grad('joint', log=log)
    sender.choose_grad('on')
    receiver.choose_grad('on')

    log('\tSystem configuration: semiosis')


def semiosis_classifier(sender_percept, sender, receiver, log=print):
    sender_percept.model_multi.train() # only optimize fc
    sender.eval()
    receiver.eval()

    sender_percept.choose_grad('last_only', log=log)
    sender.choose_grad('off')
    receiver.choose_grad('off')

    log('\tSystem configuration: last layer optim')


def semiotic_social_optimizers(state, sender_percept, sender,
                               receiver_percept, receiver):
    # optimizers define
    # ==================================
    # simulate agents condition
    agents_only(sender_percept, sender, receiver, log=log.debug)

    sender_params_to_update = []
    for name,param in sender.named_parameters():
        if param.requires_grad == True:
            sender_params_to_update.append(param)

    recv_params_to_update = []
    for name,param in receiver.named_parameters():
        if param.requires_grad == True:
            recv_params_to_update.append(param)

    static_optimizer = torch.optim.Adam([
        {'params': sender_params_to_update,
         'lr': state['sender_lr']},
        {'params': recv_params_to_update,
         'lr': state['receiver_lr']}
    ])

    if len(state['semiotic_sgd_epochs']):
        for model in [sender_percept,  sender, receiver, receiver]:
            disable_parameter_requires_grad(model)

        # simulate semiosis condition
        semiosis_joint(sender_percept, sender, receiver, log=log.debug)

        sender_params_to_update = []
        for name,param in sender.named_parameters():
            if param.requires_grad == True:
                sender_params_to_update.append(param)

        recv_params_to_update = []
        for name,param in receiver.named_parameters():
            if param.requires_grad == True:
                recv_params_to_update.append(param)

        semiotic_optimizer_specs = \
        [
            {'params': sender_params_to_update,
             'lr': state['sender_lr']},
            {'params': recv_params_to_update,
             'lr': state['receiver_lr']}
        ]
        if state['approach'] == 'proto':
            semiotic_optimizer_specs.extend([
                {'params': sender_percept.model.features.parameters(),
                 'lr': state['features_lr'],
                 'weight_decay': 1e-3},
                {'params': sender_percept.model.add_on_layers.parameters(),
                 'lr': state['add_on_layers_lr'],
                 'weight_decay': 1e-3},
                {'params': sender_percept.model.prototype_vectors,
                 'lr': state['prototype_vectors_lr']}
            ])
            classifier_optimizer_specs = [
                {'params': sender_percept.model.last_layer.parameters(),
                 'lr': state['last_layer_lr']}
            ]
        else:
            semiotic_optimizer_specs.append(
                {'params': sender_percept.model.base_model.parameters(),
                 'lr': state['features_lr'],
                 'weight_decay': 1e-3},
            )
            classifier_optimizer_specs = [
                {'params': sender_percept.model.classifier.parameters(),
                 'lr': state['last_layer_lr']}
            ]

        classifier_optimizer = torch.optim.Adam(classifier_optimizer_specs)
        semiotic_optimizer = torch.optim.Adam(semiotic_optimizer_specs)
        # joint_lr_scheduler = torch.optim.lr_scheduler.StepLR(semiotic_optimizer, step_size=10, gamma=0.1)

        return static_optimizer, semiotic_optimizer, classifier_optimizer
    else:
        return static_optimizer, None, None