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
|
#!/usr/bin/python
# Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import numpy as np
import sys
import json
import struct
def printUsage(progname):
print("%s <.json>" % (progname))
print(" This program extracts weight and bias values in TFLite format [N,H,W,C]")
print(" to .npy files in ACL format [N,C,H,W]")
print(" .npy filenames is set according to the layer's name")
if len(sys.argv) < 2:
printUsage(sys.argv[0])
exit()
filename = sys.argv[1]
f = open(filename)
j = json.loads(f.read())
tensors = j['subgraphs'][0]['tensors']
buffer_name_map = {}
for t in tensors:
if 'buffer' in t:
if t['buffer'] in buffer_name_map:
print 'find conflict!!'
print t
print buffer_name_map
comps = t['name'].split('/')
names = []
if len(comps) > 1 and comps[0] == comps[1]:
names = comps[2:]
else:
names = comps[1:]
layername = '_'.join(names)
shape = t['shape']
buffer_name_map[t['buffer']] = {'name': layername, "shape": shape}
for i in range(len(j['buffers'])):
b = j['buffers'][i]
if 'data' in b:
if i not in buffer_name_map:
print "buffer %d is not found in buffer_name_map. skip printing the buffer..."
continue
filename = "%s.npy" % (buffer_name_map[i]['name'])
shape = buffer_name_map[i]['shape']
buf = struct.pack('%sB' % len(b['data']), *b['data'])
elem_size = 1
for s in shape:
elem_size *= s
l = struct.unpack('%sf' % elem_size, buf)
n = np.array(l, dtype='f')
n = n.reshape(shape)
if len(shape) == 4:
# [N,H,W,C] -> [N,C,H,W]
n = np.rollaxis(n, 3, 1)
elif len(shape) == 3:
# [H,W,C] -> [C,H,W]
n = np.rollaxis(n, 2, 0)
elif len(shape) == 1:
pass
else:
print "Undefined length: conversion skipped. shape=", shape
#print shape, filename, n.shape
np.save(filename, n)
print "Done."
|
