/
optimizer.py
163 lines (139 loc) · 6.17 KB
/
optimizer.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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
from collections import abc as container_abcs
from collections import defaultdict
from copy import deepcopy
from itertools import chain
import torch
from ..functional import create_general_qmap, init_lpmm_generator
from ..utils import get_rank
from ..config import get_config
compression_time = 0
class LowBitOptimizer(torch.optim.Optimizer):
def __init__(self, params, defaults, config):
super(LowBitOptimizer, self).__init__(params, defaults)
# init lpmm generator
if torch.distributed.is_initialized():
seed = torch.randint(1 << 31, size=[], device=torch.device('cuda'))
torch.distributed.broadcast(seed, src=0)
init_lpmm_generator(get_rank(), seed.item()) # no stochastic rounding
self.qconfig = get_config(config)
self.override_q_enable = {}
self.qmaps = {}
def override_quantize_enable(self, module, param_name, enable):
p = getattr(module, param_name)
assert p is not None
assert isinstance(p, torch.Tensor) or isinstance(p, torch.Parameter)
if len(self.state[p]) != 0:
raise ValueError("overriding enabling of quantized parameters is prohibited after state initialization.")
self.override_q_enable[id(p)] = enable
def init_qstate(self, p, state_name):
state = self.state[p]
field = f"{state_name}_qstate"
state[field] = {
"enable": True,
"overhead": dict(),
"qmap": None,
}
subconfig = self.get_subqconfig(state_name)
state[field][
"enable"
] = _get_qenable_fn(p, subconfig.ENABLE, subconfig.THRESHOLD)
md = self.get_qmetadata_by_state_name(state_name)
qmap_key = (md['quant_type'], md['b'], md['signed'])
if qmap_key not in self.qmaps:
self.qmaps[qmap_key] = create_general_qmap(*qmap_key)
self.qmaps[qmap_key] = self.qmaps[qmap_key].to(p.device)
state[field]["qmap"] = self.qmaps[qmap_key]
def get_qmetadata_by_state_name(self, optimizer_state_name):
subconfig = self.get_subqconfig(optimizer_state_name)
md = dict(
b=subconfig.BITS,
scale_type=subconfig.SCALE_TYPE.DEFAULT,
quant_type=subconfig.QUANT_TYPE.DEFAULT,
round_type=subconfig.ROUND_TYPE,
gp_sz=subconfig.GROUP_SIZE,
signed=subconfig.SIGNED,
)
return md
def state_dict(self):
state_dict = super().state_dict()
state_dict['qconfig'] = self.qconfig
return state_dict
def load_state_dict(self, state_dict):
r"""Loads the optimizer state.
Args:
state_dict (dict): optimizer state. Should be an object returned
from a call to :meth:`state_dict`.
"""
self.qconfig = state_dict['qconfig']
# deepcopy, to be consistent with module API
state_dict = deepcopy(state_dict)
# Validate the state_dict
groups = self.param_groups
saved_groups = state_dict['param_groups']
if len(groups) != len(saved_groups):
raise ValueError("loaded state dict has a different number of "
"parameter groups")
param_lens = (len(g['params']) for g in groups)
saved_lens = (len(g['params']) for g in saved_groups)
if any(p_len != s_len for p_len, s_len in zip(param_lens, saved_lens)):
raise ValueError("loaded state dict contains a parameter group "
"that doesn't match the size of optimizer's group")
# Update the state
id_map = dict(zip(chain.from_iterable((g['params'] for g in saved_groups)),
chain.from_iterable((g['params'] for g in groups))))
def cast(param, value, key=None):
r"""Make a deep copy of value, casting all tensors to device of param."""
if isinstance(value, torch.Tensor):
# Floating-point types are a bit special here. They are the only ones
# that are assumed to always match the type of params.
# Make sure state['step'] is not casted https://github.com/pytorch/pytorch/issues/74424
if (key != "step"):
if param.is_floating_point() and value.dtype != torch.int8:
value = value.to(param.dtype)
value = value.to(param.device)
return value
elif isinstance(value, dict):
return {k: cast(param, v, key=k) for k, v in value.items()}
elif isinstance(value, container_abcs.Iterable):
return type(value)(cast(param, v) for v in value)
else:
return value
# Copy state assigned to params (and cast tensors to appropriate types).
# State that is not assigned to params is copied as is (needed for
# backward compatibility).
state = defaultdict(dict)
for k, v in state_dict['state'].items():
if k in id_map:
param = id_map[k]
state[param] = cast(param, v)
else:
state[k] = v
# Update parameter groups, setting their 'params' value
def update_group(group, new_group):
new_group['params'] = group['params']
return new_group
param_groups = [
update_group(g, ng) for g, ng in zip(groups, saved_groups)]
self.__setstate__({'state': state, 'param_groups': param_groups})
@torch.no_grad()
def step(self, closure=None):
r"""Performs a single optimization step with quantization.
Args:
closure (callable, optional): A closure that reevaluates the model
and returns the loss.
"""
raise NotImplementedError(
'The step method needs overriding'
)
def get_subqconfig(self, optimizer_state_name):
raise NotImplementedError(
'The get_subconfig method needs overriding'
)
def _get_qenable_fn(p, prior_enable, th):
if not prior_enable:
return False
# if p.dim() < 2:
# return False
if th is not None and p.numel() <= th:
return False
return True