Python
matplotlib
https://matplotlib.org/contents.html
numpy
pytorch
英文文檔:https://pytorch.org/docs/stable/index.html
中文文檔:https://pytorch-cn.readthedocs.io/zh/latest/package_references/torch/
scikit-image
https://scikit-image.org/docs/stable/
pickle
https://docs.python.org/3.5/library/pickle.html
easydict(字典解析庫)
https://pypi.org/project/easydict/1.2/
os(操作系統接口模塊)
https://docs.python.org/3/library/os.html?highlight=os#module-os
sys
https://docs.python.org/3/library/sys.html?highlight=sys#module-sys
argparse(解析命令行)
https://docs.python.org/3/library/argparse.html?highlight=argparse#module-argparse
fire(命令行生成工具)
https://github.com/google/python-fire/blob/master/docs/guide.md
logging(日誌)
https://docs.python.org/3.6/howto/logging.html
tensorboardX(網絡可視化)
https://tensorboardx.readthedocs.io/en/latest/tutorial.html
tqdm(進度條工具)
https://pypi.org/project/tqdm/
glob(文件搜索庫)
https://docs.python.org/3/library/glob.html?highlight=glob#module-glob
re(正則化)
https://docs.python.org/3/library/re.html
io
https://docs.python.org/3/library/io.html?highlight=io#module-io
time
https://docs.python.org/3/library/time.html?highlight=time#module-time
datitime
https://docs.python.org/3/library/datetime.html?highlight=datetime#module-datetime
numba(Python加速庫)
import numpy as np
import numba
from numba import jit
print(numba.__version__)
@jit(nopython=True)
'''
The nopython=True option requires that the function be fully compiled
'''
def go_fast(a): # Function is compiled to machine code when called the first time
trace = 0
# assuming square input matrix
for i in range(a.shape[0]): # Numba likes loops
trace += np.tanh(a[i, i]) # Numba likes NumPy functions
return a + trace # Numba likes NumPy broadcasting
x = np.arange(100).reshape(10, 10)
go_fast(x)
go_fast(2*x)
%timeit go_fast(x)
np.testing.assert_array_equal(go_fast(x), go_fast.py_func(x))
%timeit go_fast.py_func(x)
def go_numpy(a):
return a + np.tanh(np.diagonal(a)).sum()
np.testing.assert_array_equal(go_numpy(x), go_fast(x))
%timeit go_numpy(x)