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
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
|
--- Implements a sequence-level chopped and shuffled buffer which
-- shall be used for cyclic NNs.
-- @author Qi Liu <liuq901@163.com>
--- The class for a sequence-level chopped and shuffled buffer which
-- shall be used for cyclic NNs.
-- @type nerv.SeqBuffer
local SeqBuffer = nerv.class('nerv.SeqBuffer', 'nerv.DataBuffer')
--- The constructor.
-- @param global_conf a table describing the computation state and providing
-- with some global settings
--
-- The following fields in `global_conf` will be used:
--
-- * `use_cpu`: whether to provide with the chunks/"mini-batches" stored in the
-- main memory on invocation of `get_data()`
-- * `mmat_type`: the class used for creating matrices in CPU computation
-- * `cumat_type` (if `use_cpu = false`): the class used for creating matrices
-- in GPU computation
--
-- @param buffer_conf a table providing with settings dedicated for the buffer.
-- Available fields includes:
--
-- * `readers`: an array of `nerv.DataReader` instances specifying the
-- readers used to read data
-- * `batch_size`: the number of rows for each batch matrix
-- * `chunk_size`: the length of the BPTT context (number of batch
-- matrices to provide upon each invocation of `get_data()`)
-- * `buffer_size`: the number of frames to be buffered and shuffled at once (shuffle
-- in the sequence level, not in the frame level)
-- * `randomize`: shuffle the buffer after filled if true
-- * `nn_act_default`: the default value to fill into the "holes" (non-data
-- frames)
function SeqBuffer:__init(global_conf, buffer_conf)
self.gconf = global_conf
self.batch_size = buffer_conf.batch_size
self.chunk_size = buffer_conf.chunk_size
self.buffer_size = buffer_conf.buffer_size
self.randomize = buffer_conf.randomize
self.readers = {}
for _, v in ipairs(buffer_conf.readers) do
table.insert(self.readers, v.reader)
end
self.nn_act_default = buffer_conf.nn_act_default
if self.nn_act_default == nil then
self.nn_act_default = 0
end
self.mat_type = self.gconf.mmat_type
self.queue = {}
self.head = 1
self.tail = 0
self.offset = 1
self.buffer = {}
self.length = {}
self.index = {}
self.complete = false
end
function SeqBuffer:new_mini_batch()
local res = {}
res.data = {}
res.new_seq = {}
res.seq_length = {}
for i = 1, self.batch_size do
res.seq_length[i] = 0
end
return res
end
local function random_shuffle(a)
for i = #a, 2, -1 do
local j = math.random(i)
a[i], a[j] = a[j], a[i]
end
end
function SeqBuffer:fill_buffer()
if self.complete then
return false
end
local t = os.clock()
self.buffer = {}
self.length = {}
local size = 0
while size < self.buffer_size do
local drow = nil
local data = {}
for i = 1, #self.readers do
local tmp = self.readers[i]:get_data()
if tmp == nil then
self.complete = true
break
end
for id, d in pairs(tmp) do
if drow == nil then
drow = d:nrow()
elseif d:nrow() ~= drow then
nerv.error('readers provides with inconsistent rows of data')
end
data[id] = d
end
end
if self.complete then
break
end
size = size + drow
table.insert(self.buffer, data)
table.insert(self.length, drow)
end
self.index = {}
for i = 1, #self.buffer do
self.index[i] = i
end
if self.randomize then
random_shuffle(self.index)
end
self.offset = 1
collectgarbage('collect')
nerv.info('%.3fs to fill the buffer', os.clock() - t)
return #self.buffer > 0
end
function SeqBuffer:get_buffered_data()
if self.offset > #self.buffer then
if not self:fill_buffer() then
return nil
end
end
local id = self.index[self.offset]
self.offset = self.offset + 1
return self.buffer[id], self.length[id]
end
function SeqBuffer:saturate(batch)
if self.queue[self.head] ~= nil and self.queue[self.head].seq_length[batch] ~= 0 then
return true
end
local data, drow = self:get_buffered_data()
if data == nil then
return false
end
local offset = 0
local head = self.head
while offset < drow do
local last = math.min(offset + self.chunk_size, drow)
if head > self.tail then
self.tail = self.tail + 1
self.queue[self.tail] = self:new_mini_batch()
end
self.queue[head].seq_length[batch] = last - offset
if offset == 0 then
table.insert(self.queue[head].new_seq, batch)
end
local mini_batch = self.queue[head].data
for id, d in pairs(data) do
if mini_batch[id] == nil then
mini_batch[id] = {}
end
local tmp = mini_batch[id]
for i = offset + 1, last do
local chunk = i - offset
if tmp[chunk] == nil then
tmp[chunk] = self.mat_type(self.batch_size, d:ncol())
tmp[chunk]:fill(self.nn_act_default)
end
tmp[chunk]:copy_from(d, i - 1, i, batch - 1)
end
end
head = head + 1
offset = last
end
return true
end
--- Get a batch group from the buffer.
-- See `nerv.DataBuffer` for reference.
function SeqBuffer:get_data()
local has_data = false
for i = 1, self.batch_size do
if self:saturate(i) then
has_data = true
end
end
if not has_data then
return nil
end
local res = self.queue[self.head]
self.queue[self.head] = nil
self.head = self.head + 1
if not self.gconf.use_cpu then
for id, d in pairs(res.data) do
for i = 1, #d do
d[i] = self.gconf.cumat_type.new_from_host(d[i])
end
end
end
return res
end
|