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
|
# Copyright (c) Facebook, Inc. and its affiliates.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from bitsandbytes.optim.optimizer import Optimizer1State
class Adagrad(Optimizer1State):
def __init__(
self,
params,
lr=1e-2,
lr_decay=0,
weight_decay=0,
initial_accumulator_value=0,
eps=1e-10,
optim_bits=32,
args=None,
min_8bit_size=4096,
percentile_clipping=100,
block_wise=True,
):
if not 0.0 <= lr:
raise ValueError("Invalid learning rate: {}".format(lr))
if not 0.0 <= weight_decay:
raise ValueError(
"Invalid weight_decay value: {}".format(weight_decay)
)
if not 0.0 <= eps:
raise ValueError("Invalid epsilon value: {}".format(eps))
if initial_accumulator_value != 0.0:
raise ValueError("Initial accumulator value != 0.0 not supported!")
if lr_decay != 0.0:
raise ValueError("Lr Decay != 0.0 not supported!")
super(Adagrad, self).__init__(
"adagrad",
params,
lr,
(0.0, 0.0),
eps,
weight_decay,
optim_bits,
args,
min_8bit_size,
percentile_clipping,
block_wise,
)
class Adagrad8bit(Optimizer1State):
def __init__(
self,
params,
lr=1e-2,
lr_decay=0,
weight_decay=0,
initial_accumulator_value=0,
eps=1e-10,
optim_bits=8,
args=None,
min_8bit_size=4096,
percentile_clipping=100,
block_wise=True,
):
if not 0.0 <= lr:
raise ValueError("Invalid learning rate: {}".format(lr))
if not 0.0 <= weight_decay:
raise ValueError(
"Invalid weight_decay value: {}".format(weight_decay)
)
if not 0.0 <= eps:
raise ValueError("Invalid epsilon value: {}".format(eps))
if initial_accumulator_value != 0.0:
raise ValueError("Initial accumulator value != 0.0 not supported!")
if lr_decay != 0.0:
raise ValueError("Lr Decay != 0.0 not supported!")
assert block_wise
super(Adagrad8bit, self).__init__(
"adagrad",
params,
lr,
(0.0, 0.0),
eps,
weight_decay,
8,
args,
min_8bit_size,
percentile_clipping,
block_wise,
)
class Adagrad32bit(Optimizer1State):
def __init__(
self,
params,
lr=1e-2,
lr_decay=0,
weight_decay=0,
initial_accumulator_value=0,
eps=1e-10,
optim_bits=32,
args=None,
min_8bit_size=4096,
percentile_clipping=100,
block_wise=True,
):
if not 0.0 <= lr:
raise ValueError("Invalid learning rate: {}".format(lr))
if not 0.0 <= weight_decay:
raise ValueError(
"Invalid weight_decay value: {}".format(weight_decay)
)
if not 0.0 <= eps:
raise ValueError("Invalid epsilon value: {}".format(eps))
if initial_accumulator_value != 0.0:
raise ValueError("Initial accumulator value != 0.0 not supported!")
if lr_decay != 0.0:
raise ValueError("Lr Decay != 0.0 not supported!")
super(Adagrad32bit, self).__init__(
"adagrad",
params,
lr,
(0.0, 0.0),
eps,
weight_decay,
32,
args,
min_8bit_size,
percentile_clipping,
block_wise,
)
|
