Reference: Format Specification Mini-Language.
Quick demo of desired behavior and motivation:
>>> x = np.array([100.002, 1.2])
# use the same data augmentation scheme as the FlowNet2 paper | |
layer { | |
name: "img0" | |
type: "CustomData" | |
top: "img0" | |
top: "img1" | |
top: "flow_gt" | |
top: "aux" | |
include { | |
phase: TRAIN |
Reference: Format Specification Mini-Language.
Quick demo of desired behavior and motivation:
>>> x = np.array([100.002, 1.2])
from __future__ import division, print_function, absolute_import | |
import os | |
import struct | |
from array import array | |
import numpy as np | |
def load_mnist(section="training", offset=0, count=None, ret='xy', | |
x_dtype=np.float64, y_dtype=np.int64, path=None): | |
""" |
from __future__ import division, print_function, absolute_import | |
import os | |
import struct | |
from array import array | |
import numpy as np | |
def load_mnist(section="training", offset=0, count=None, ret='xy', | |
x_dtype=np.float64, y_dtype=np.int64, path=None): | |
""" |
# Author: Nelle Varoquaux, Andrew Tulloch | |
# Uses the pool adjacent violators algorithm (PAVA), with the | |
# enhancement of searching for the longest decreasing subsequence to | |
# pool at each step. | |
import numpy as np | |
cimport numpy as np | |
cimport cython |