forked from karin0018/FA-DAT
-
Notifications
You must be signed in to change notification settings - Fork 0
/
setup.py
288 lines (245 loc) · 8.35 KB
/
setup.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
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
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
#!/usr/bin/env python3
# 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.
import os
import subprocess
import site
import sys
site.ENABLE_USER_SITE = "--user" in sys.argv[1:]
from setuptools import Extension, find_packages, setup
if sys.version_info < (3, 6):
sys.exit("Sorry, Python >= 3.6 is required for fairseq.")
def write_version_py():
with open(os.path.join("fairseq", "version.txt")) as f:
version = f.read().strip()
# append latest commit hash to version string
try:
sha = (
subprocess.check_output(["git", "rev-parse", "HEAD"])
.decode("ascii")
.strip()
)
version += "+" + sha[:7]
except Exception:
pass
# write version info to fairseq/version.py
with open(os.path.join("fairseq", "version.py"), "w") as f:
f.write('__version__ = "{}"\n'.format(version))
return version
version = write_version_py()
with open("README.md") as f:
readme = f.read()
if sys.platform == "darwin":
extra_compile_args = ["-stdlib=libc++", "-O3"]
else:
extra_compile_args = ["-std=c++11", "-O3"]
class NumpyExtension(Extension):
"""Source: https://stackoverflow.com/a/54128391"""
def __init__(self, *args, **kwargs):
self.__include_dirs = []
super().__init__(*args, **kwargs)
@property
def include_dirs(self):
import numpy
return self.__include_dirs + [numpy.get_include()]
@include_dirs.setter
def include_dirs(self, dirs):
self.__include_dirs = dirs
extensions = [
Extension(
"fairseq.libbleu",
sources=[
"fairseq/clib/libbleu/libbleu.cpp",
"fairseq/clib/libbleu/module.cpp",
],
extra_compile_args=extra_compile_args,
),
NumpyExtension(
"fairseq.data.data_utils_fast",
sources=["fairseq/data/data_utils_fast.pyx"],
language="c++",
extra_compile_args=extra_compile_args,
),
NumpyExtension(
"fairseq.data.token_block_utils_fast",
sources=["fairseq/data/token_block_utils_fast.pyx"],
language="c++",
extra_compile_args=extra_compile_args,
),
]
cmdclass = {}
try:
# torch is not available when generating docs
from torch.utils import cpp_extension
# extensions.extend(
# [
# cpp_extension.CppExtension(
# "fairseq.libbase",
# sources=[
# "fairseq/clib/libbase/balanced_assignment.cpp",
# ],
# )
# ]
# )
# extensions.extend(
# [
# cpp_extension.CppExtension(
# "fairseq.libnat",
# sources=[
# "fairseq/clib/libnat/edit_dist.cpp",
# ],
# ),
# cpp_extension.CppExtension(
# "alignment_train_cpu_binding",
# sources=[
# "examples/operators/alignment_train_cpu.cpp",
# ],
# ),
# ]
# )
if "CUDA_HOME" in os.environ:
pass
# extensions.extend(
# [
# cpp_extension.CppExtension(
# "fairseq.libnat_cuda",
# sources=[
# "fairseq/clib/libnat_cuda/edit_dist.cu",
# "fairseq/clib/libnat_cuda/binding.cpp",
# ],
# ),
# cpp_extension.CppExtension(
# "fairseq.ngram_repeat_block_cuda",
# sources=[
# "fairseq/clib/cuda/ngram_repeat_block_cuda.cpp",
# "fairseq/clib/cuda/ngram_repeat_block_cuda_kernel.cu",
# ],
# ),
# cpp_extension.CppExtension(
# "alignment_train_cuda_binding",
# sources=[
# "examples/operators/alignment_train_kernel.cu",
# "examples/operators/alignment_train_cuda.cpp",
# ],
# ),
# ]
# )
cmdclass["build_ext"] = cpp_extension.BuildExtension
except ImportError:
pass
if "READTHEDOCS" in os.environ:
# don't build extensions when generating docs
extensions = []
if "build_ext" in cmdclass:
del cmdclass["build_ext"]
# use CPU build of PyTorch
dependency_links = [
"https://download.pytorch.org/whl/cpu/torch-1.7.0%2Bcpu-cp36-cp36m-linux_x86_64.whl"
]
else:
dependency_links = []
if "clean" in sys.argv[1:]:
# Source: https://bit.ly/2NLVsgE
print("deleting Cython files...")
import subprocess
subprocess.run(
["rm -f fairseq/*.so fairseq/**/*.so fairseq/*.pyd fairseq/**/*.pyd"],
shell=True,
)
extra_packages = []
if os.path.exists(os.path.join("fairseq", "model_parallel", "megatron", "mpu")):
extra_packages.append("fairseq.model_parallel.megatron.mpu")
def do_setup(package_data):
setup(
name="fairseq",
version=version,
description="Facebook AI Research Sequence-to-Sequence Toolkit",
url="https://github.com/pytorch/fairseq",
classifiers=[
"Intended Audience :: Science/Research",
"License :: OSI Approved :: MIT License",
"Programming Language :: Python :: 3.6",
"Programming Language :: Python :: 3.7",
"Programming Language :: Python :: 3.8",
"Topic :: Scientific/Engineering :: Artificial Intelligence",
],
long_description=readme,
long_description_content_type="text/markdown",
setup_requires=[
"cython",
'numpy<1.20.0; python_version<"3.7"',
'numpy; python_version>="3.7"',
"setuptools>=18.0",
],
install_requires=[
"cffi",
"cython",
'dataclasses; python_version<"3.7"',
"hydra-core>=1.0.7,<1.1",
"omegaconf<2.1",
'numpy<1.20.0; python_version<"3.7"',
'numpy; python_version>="3.7"',
"regex",
"sacrebleu==1.5.1",
"sacrebleu[ja]",
"tqdm",
"bitarray",
"ninja"
],
dependency_links=dependency_links,
packages=find_packages(
exclude=[
"examples",
"examples.*",
"scripts",
"scripts.*",
"tests",
"tests.*",
]
)
+ extra_packages,
package_data=package_data,
ext_modules=extensions,
test_suite="tests",
entry_points={
"console_scripts": [
"fairseq-eval-lm = fairseq_cli.eval_lm:cli_main",
"fairseq-generate = fairseq_cli.generate:cli_main",
"fairseq-hydra-train = fairseq_cli.hydra_train:cli_main",
"fairseq-interactive = fairseq_cli.interactive:cli_main",
"fairseq-preprocess = fairseq_cli.preprocess:cli_main",
"fairseq-score = fairseq_cli.score:cli_main",
"fairseq-train = fairseq_cli.train:cli_main",
"fairseq-validate = fairseq_cli.validate:cli_main",
],
},
cmdclass=cmdclass,
zip_safe=False,
)
def get_files(path, relative_to="fairseq"):
all_files = []
for root, _dirs, files in os.walk(path, followlinks=True):
root = os.path.relpath(root, relative_to)
for file in files:
if file.endswith(".pyc"):
continue
all_files.append(os.path.join(root, file))
return all_files
if __name__ == "__main__":
try:
# symlink examples into fairseq package so package_data accepts them
fairseq_examples = os.path.join("fairseq", "examples")
if "build_ext" not in sys.argv[1:] and not os.path.exists(fairseq_examples):
os.symlink(os.path.join("..", "examples"), fairseq_examples)
package_data = {
"fairseq": (
get_files(fairseq_examples)
+ get_files(os.path.join("fairseq", "config"))
)
}
do_setup(package_data)
finally:
if "build_ext" not in sys.argv[1:] and os.path.islink(fairseq_examples):
os.unlink(fairseq_examples)