b****e
发帖数: 1275 | 1 is there an oracle function that does what the perl localtime does?
ie. translate the epoch date (long int) to a date string like mm/dd/yyyy?
thanks |
|
b****e
发帖数: 1275 | 2 no no no.. i am not asking for that.. i need a function that
translates epoch time (sec from jan 1 1970) to MMDDYYYY kind
of format. to_char doesn't work
thanks |
|
b****e
发帖数: 1275 | 3 here's a problem i have.. a stupid software vendor P stores
their date format in unix epoch (seconds from 1/1/1970) and
all other applications we have can only handle standard
oracle dates (MM/DD/YYYY). so in order to make them talk to
each other i have to add a column (oracle_date) to the table
used by software P. and every time that software does an
insert or update i want to make sure my additional column
gets updated as well..
so i wanted to use a database trigger on this table..
CREATE TRIG |
|
|
s******v
发帖数: 4495 | 5 but ipv4 connectivity with label is carried by IGP.
when I did show ip bgp
PE1#sh ip bgp 44.100.100.2
BGP routing table entry for 44.100.100.2/32, version 5
Paths: (1 available, best #1, table default, RIB-failure(17))
Not advertised to any peer
Refresh Epoch 1
Local
44.100.100.2 (metric 2) from 44.100.100.2 (44.100.100.22)
Origin incomplete, metric 0, localpref 100, valid, internal, best
rx pathid: 0, tx pathid: 0x0
this ipv4 prefix is even marked as RIB f... 阅读全帖 |
|
F***Q
发帖数: 6599 | 6
it's all database operations.
load-on-page-bottom is not so different from paging by clicking numbered
links. the key is the URL.
if you use developer tools in chrome/firefox, when the loading event happens
, you should see the URL that ajax is trying to retrieve. many cases, you
can decipher the parameters - likely you will see something to control the
starting time (as epoch time), end-time or record length. with that URL
understood, you can pretty much load anything you want by typing it in ... 阅读全帖 |
|
p**s
发帖数: 2707 | 7 convert to Julian day number
or
convert to Epoch time |
|
w***n
发帖数: 1137 | 8 看到一个程序， 里面有这么一句。
#include
#include
#include "errors.h"
typedef struct alarm_tag {
struct alarm_tag *link;
int seconds;
time_t time; /* seconds from EPOCH */
char message[64];
} alarm_t;
在主程序力有这么一句
alarm->time = time (NULL) + alarm->seconds;
以前没有见过这种用法，这句到底是什么意思？ |
|
T********i
发帖数: 2416 | 9 deadline timer是等到某个时间以后返回。
duration timer是一个指定的时间间隔。
deadline time depends on system timer. System timer can be modified by users
, or adjusted by NTP etc. Especially for embedded system, the initial time
will be unix epoch until the NTP client gets the current timer.
Linux IPC很多同步都基于deadline timer。 |
|
|
b***d
发帖数: 288 | 11 刚开始学习，两个问题想学习一下，
1. “制约CNN的不是硬件而是sigmoid函数”
sigmoid函数怎么制约了CNN？
2. 网络里很多地方引入了随机数，比如在DBN中的初始场和激活函数，
为什么一定要引入随机？而实际计算过程中，epoch,batch循环时这个貌似随机数 并
不是真正的随机（每次循环时都是相同的）。 |
|
m****o
发帖数: 182 | 12 应该是可以的。transfer learning做起来稍微比较麻烦一点。下面的程序极有可能不
能跑通，我直接在网页上写出来的。但是你可以知道我的意思。只获取数字和字母模型
的隐含层信息，然后把两个的最后两层concat在一起和输出层做一个全联接。训练的时
候冻结所有隐含层的权值更新。这样的话你只是训练最后一层即输出层的权值，速度应
该会快很多。如果效果不理想，可以在输出层之前再加上一层银行层。
import tensorflow as tf
from tensorflow.contrib.layers import fully_connected,
n_in = 28 * 28 # assuming the size of mnist
n_out = 10 + 26 # 0-9 + 'a'-'z'
n_hidden = 100 # assuming both models have two hidden layers with size = 100
with tf.Graph().as_default():
X = tf.placeholder(tf.float32, shape=... 阅读全帖 |
|
w*****h
发帖数: 423 | 13 正在测试CNN，epoch还没训练完，不过目前的accuracy已经达到70%，
这还是在没有做ensemble的情况下， 太赞了。
再请教wdong和楼上几位，有必要再上RNN吗？ |
|
w*****h
发帖数: 423 | 14 CNN
可能有点误解，我是说CNN训练的轮数还没达到设定的epoch值， 它还在训练中，但目
前的model已经有70% accuracy了
xgboost我也试过了，和NN差不多 |
|
w***g
发帖数: 5958 | 15 如果是几十上百个词的短文本，我估计rnn不会有提高（NLP专家请指正）。你有时间还
不如试试data augmentation, 比如随机增减几个词，同义词替换啥的。不过你现在最
重要的是establish问题复杂度，就是上次说的对比不同人的区别，这样知道什么时候
停止优化。不然老板拍脑袋定指标，终有一天会把你逼死。
：正在测试CNN，epoch还没训练完，不过目前的accuracy已经达到70%，
： |
|
w*****h
发帖数: 423 | 16 正在测试CNN，epoch还没训练完，不过目前的accuracy已经达到70%，
这还是在没有做ensemble的情况下， 太赞了。
再请教wdong和楼上几位，有必要再上RNN吗？ |
|
w*****h
发帖数: 423 | 17 CNN
可能有点误解，我是说CNN训练的轮数还没达到设定的epoch值， 它还在训练中，但目
前的model已经有70% accuracy了
xgboost我也试过了，和NN差不多 |
|
w***g
发帖数: 5958 | 18 如果是几十上百个词的短文本，我估计rnn不会有提高（NLP专家请指正）。你有时间还
不如试试data augmentation, 比如随机增减几个词，同义词替换啥的。不过你现在最
重要的是establish问题复杂度，就是上次说的对比不同人的区别，这样知道什么时候
停止优化。不然老板拍脑袋定指标，终有一天会把你逼死。
：正在测试CNN，epoch还没训练完，不过目前的accuracy已经达到70%，
： |
|
g*******u
发帖数: 3948 | 19 这个max steps 参数 什么含义啊？ 和常用的 epochs 什么联系呢？
看了官网 貌似也没说清楚
ps tf 的 文档 真是 难读啊 感觉 就是为了交叉 吗足够多的字 没太想 到底读者
看着清楚否 |
|
i*****9
发帖数: 3157 | 20 就是一共训练 max steps 个 batch 就结束 和训练集大小无关
：这个max steps 参数 什么含义啊？ 和常用的 epochs 什么联系呢？
：看了官网 貌似也没说清楚 |
|
w***g
发帖数: 5958 | 21 如果你有100G数据，想在每个epoch reshuffle，估计就独我一家可以支持了。
虽然其实reshuffle了也是然并卵。但他们那么建议是因为软件设计无法支持
reshuffle。Design decision不一样，代码没什么困难的。 |
|
x**********i
发帖数: 658 | 22 神经网络有没有个大约general的共识，比如batch size越大越好或者某个值，epoch越
多越好？filter越多越好？
还是这些参数必须根据具体model来调整
：我觉得我能理解这个现象，我猜如果你不shuffle，应该不会有这个现象
： |
|
发帖数: 1 | 23 你的第二个链接里也说了 reducing the learning rate by 10 after 150 epochs |
|
w*****r
发帖数: 197 | 24 Adam, sgd, momentum,...有啥讲究没?
还是，评价某一个改动对预测性能的影响，一般跑几个epoch就能看出来？ |
|
w***g
发帖数: 5958 | 25 如果几个epoch明显看到了提高，最后往往也有提高。
这个真的是炼金术。至少我目前看不到什么道理可言。
搞个贵点的设备，快速iterate是硬道理。 |
|
x****u
发帖数: 44466 | 26 生物的确没有bp，但也做不到200个epoch进化出imagenet视觉来
你希望网络结构迅速进化出权重就离不开bp |
|
f*r
发帖数: 3968 | 27 LINEAR彗星（C2000 WM1）粉墨登场
译：Shea
在11月份，LINEAR彗星会穿越英仙座，用双筒镜也能轻松的找到它。在月底，说不定用肉
眼也能勉强看见它。南半球的观测者在12月份会有一个比较好的观测条件。
尽管这颗彗星在去年12月份被发现时只有18等，但它自那之后便开始慢慢的增亮。这儿的
星表是基于Syuichi Nakano提供的最新轨道根数计算的结果，其中给出了表中所给日期世
界时0时这颗彗星的赤经和赤纬，它和太阳的角距，目视亮度由John Bortle提供，以及它
所在的星座。
C/2000 WM1 (LINEAR) 轨道根数
过近日点时间 T：2002 Jan. 22.6747 TT
轨道偏心率 e：1.000281
近日点距 q: 0.555366 a.u.
近日点角距 ω：276.7703
轨道升交点黄经 Ω：237.8964
轨道面与黄道面的交角 i：72.5514
Epoch 2001 Oct. 18.0 TT = JDT 2452200.5
T 2002 Jan. 22.6747 TT Marsden
q 0.555366 (2000.0) P Q
z |
|
s*****t
发帖数: 1994 | 28 BEIJING -- China is about to join the hunt for dark energy. At a cosmology
workshop held here on 20 March, scientists unveiled Tianlai, or "Sound of
Heaven," a project to listen to radio emissions from deep space that may
reveal the nature of dark energy.
Dark energy is the mysterious force driving the accelerating expansion of
the universe. Although not directly observable, the phenomenon "can be
studied by observing the expansion rate of the universe at different epochs,
" says Tianlai project... 阅读全帖 |
|
s*****t
发帖数: 1994 | 29 本篇的主人公就是曾经活跃在本版的网友哟。
epochs,
waves |
|
s*******n
发帖数: 1474 | 30 恭喜xuelei拉到经费阿
epochs,
waves |
|
g*********d
发帖数: 233 | 31 Genetic history of an archaic hominin group from Denisova Cave in
Siberia
David Reich, Richard E. Green, Martin Kircher, Johannes Krause,
Nick Patterson, Eric Y. Durand, Bence Viola, Adrian W. Briggs, Udo
Stenzel, Philip L. F. Johnson, Tomislav Maricic, Jeffrey M. Good,
Tomas Marques-Bonet, Can Alkan, Qiaomei Fu, Swapan Mallick, Heng
Li, Matthias Meyer, Evan E. Eichler, Mark Stoneking, Michael
Richards, Sahra Talamo, Michael V. Shunkov, Anatoli P. Derevi... 阅读全帖 |
|
m****n
发帖数: 1066 | 32 epoch in houston
very good. |
|
|
f****g
发帖数: 5 | 34 atm000: what are you talking about? I have some close friends working for
big 4. You should know that Big 4 not just do auditing and tax, but also do
consulting in many other areas.
What is EPOCH.NET的轮子? |
|
f****g
发帖数: 5 | 35 atm000: what are you talking about? I have some close friends working for
big 4. You should know that Big 4 not just do auditing and tax, but also do
consulting in many other areas.
What is EPOCH.NET的轮子? |
|
p******e
发帖数: 1151 | 36 这是Newton相提并论的说法的出处 （wiki上找的）。
不过百年几遇是Gauss的原文， 应该比较可信。 百年几遇评价当然非常非常高，我觉
得可能也比较恰当（虽然不懂数论）。 不过和人类历史上仅有的Newton相比（千年一
遇的级别）， 就是菩萨和罗汉的区别. 即使Gauss抬爱， 估计也很难说到一块去。
E. T. Bell in his 1937 book Men of Mathematics (page 237) claims that Gauss
said "There have been but three epoch-making mathematicians, Archimedes,
Newton, and Eisenstein", and this has been widely quoted in writings about
Eisenstein. This is not a quote by Gauss, but is (a translation of) the end
of a sentence from the biography of Eisenste... 阅读全帖 |
|
s*********a
发帖数: 336 | 37 是不是有免疫缺陷的病人, 这种病人可以得N次sepsis. Epoch是时间段的意思? 这个不
太确定. |
|
d**c
发帖数: 515 | 38 先谢谢你了啊，我也看得迷迷糊糊的说
epoch，epidose貌似都是一些医学上的术语，我这个外行，看的迷迷糊糊的~
我也不知道那个sepsis到底是那种病，貌似应该是败血症，这个文章是预测 新生儿败
血症 |
|
d****y
发帖数: 53 | 39 Sure. To derive this formula only requires stat. mech. (the $\eta$ I mentioned
is just the mean occupation number for bosons/fermions) One derives the energy
density for each particle sepcies and then adds them up.
I mentioned cosmology b/c the formula you mentioned (energy
density for several particle species) often finds its application in
cosmology: e.g. when you want to consider the energy density of the
universe at a certain epoch, when there're photons, neutrinos, electrons... |
|
Y******Y
发帖数: 8753 | 40 你仔细看说明：
individual
This argument has been superseded by the id argument and is present only for
backwards compatability. A logical value indicating whether each row of
newdata represents a distinct individual (FALSE, the default), or if each
row of the data frame represents different time epochs for only one
individual (TRUE). In the former case the result will have one curve for
each row in newdata, in the latter only a single curve will be produced.
就是说，if one row per subject, then individu... 阅读全帖 |
|
|
y**********u
发帖数: 6366 | 42 好像是Craig快一点
http://www.slowtwitch.com/News/Crowie_takes_3rd_Kona_title_Chri
Alexander combined a 51:56 swim, a shockingly impressive second-best 4:24:05
bike split and a second-best 2:44:03 marathon to hit the tape in 8:03:56
with a 5:15 margin over surprise runner-up and fellow Australian Pete Jacobs
and a lucky 7:11 over pre-race favorite Andreas Raelert of Germany, last
year's runner-up in an epic duel with Chris McCormack. Alexander also broke
by 12 seconds the 15-year-old race record of 8:0... 阅读全帖 |
|
I***i
发帖数: 14557 | 43 http://www.sciencerecorder.com/news/243-million-year-old-dinosa
243 million-year-old dinosaur discovered in London museum
The world’s oldest dinosaur may have been hiding in plain sight.
Researchers at London’s Natural History Museum say an astounding 243
million-year-old dinosaur fossil has been discovered. The creature, which
has been under study for over two decades, is widely seen as the old
dinosaur ever unearthed by paleontologists.
“If the newly named Nyasasaurus parringtoni is not the ea... 阅读全帖 |
|
d******a
发帖数: 32122 | 44 https://opentreeoflife.files.wordpress.com/2013/10/pantherablog.png
第四张是最新2010年的研究成果
A 2010 study published in Molecular Phylogenetics and Evolution has given
insight into the exact evolutionary relationships of the big cats.[10] The
study reveals that the snow leopard and the tiger are sister species, while
the lion, leopard, and jaguar are more closely related to each other. The
tiger and snow leopard diverged from the ancestral big cats approximately 3.
9 Ma. The tiger then evolved into a uni... 阅读全帖 |
|
|
N***i
发帖数: 2063 | 46 Best-Rated Financial Stocks Updated June 23, 2010
http://www.nasd100.com/2010/06/bestrated-financial-stocks-updated-june-23-2010.html
Ranking | Company (Ticker) | Average Rating Value | Average Rating
1 IRSA Inversiones Representaciones (ADR) (NYSE:IRS) 1.00 Strong Buy
2 Primus Guaranty, Ltd. (NYSE:PRS) 1.00 Strong Buy
3 China Life Insurance Company Ltd. (ADR) (NYSE:LFC) 1.00 Strong Buy
4 National Bank of Greece (ADR) (NYSE:NBG) 1.00 Strong Buy
5 HDFC Bank Limited (ADR) (NYSE:HDB) 1.00 ... 阅读全帖 |
|
f*****a
发帖数: 781 | 47 你相信 Terrance Mckenna 的 Timewave Zero 不？
It matches the signature of time mathematically and soon we will run out
of time.
This time of history is truly the end of one epoch.
through a
flu is
with
then the |
|
f*****a
发帖数: 781 | 48 你相信 Terrance Mckenna 的 Timewave Zero 不？
It matches the signature of time mathematically and soon we will run out
of time.
This time of history is truly the end of one epoch.
through a
flu is
with
then the |
|
