由买买提看人间百态

topics

全部话题 - 话题: targted
(共0页)
O*******k
发帖数: 1743
1
多谢提醒,已经将DCPI和fight tgt都更新到猎兔犬里面了
以后步入targt后,别忘记用手机让猎兔犬扫一下 target的 ipad 结果,
然后用显示结果去fight就OK了,通常有货说没货的,都会最终给的
c******g
发帖数: 9273
2
targte首页上有
f****o
发帖数: 2770
3
来自主题: Medicine版 - 求助,肺癌病人.
honestly, i don't know what to tell you. i am sure you know at stage IV,
there is not much we can do.
If the cancer cell has specific mutation, there are erlotinib/gefitinib and
crizotinib that are targted therapy. Only if your relative has those
specific mutation. Those meds are expensive and only prolong survival. There
is no cure for stage IV lung cancer. 5 year survival is only 1-2%
realistic speaking, your relative should do whatever he/she wants to do
while still alive
palliative care/symp... 阅读全帖
s*********t
发帖数: 16647
4
让我猜猜,你是不是注册过citi季度性的5% offer?
这种offer是targted的,会发到很多citi cc上的
h*******n
发帖数: 19
5
来自主题: Money版 - Pls comment: Marriott targted offer
$200 Marriott gift card + 50,000 points after spending $3K in 3 months.
First year annual fee waived. Is this a good deal?
h*******n
发帖数: 19
6
来自主题: Money版 - Pls comment: Marriott targted offer
Forgot to mention: 1 free night every year (Cat 1-5) after anniversary
a*****a
发帖数: 302
7
来自主题: Money版 - Pls comment: Marriott targted offer
最高好像80000点吧,你算算30000点和200刀哪个值钱
G*8
发帖数: 13261
8
来自主题: PennySaver版 - 要不要马上冲出去打针???
附近都木有针,拔草省心了。
你家娃要打,targt有免费的,不过肯定是没货了。
l******k
发帖数: 363
9
来自主题: PennySaver版 - CVS25折收获
那个益智球是跟targt类似的吗?很好玩,可惜周围店都没有。
j**********9
发帖数: 5431
10
来自主题: Chicago版 - 男人们打猎回来
it is bigbee who is targted......
I should have gone to join you guys.
a*****a
发帖数: 1038
11
来自主题: Environmental版 - career goal 咋说?
there are only 3 career goals for any environmental consultant:
1) Technical Director
2) Project Manager
3) Client Officer
Sometimes you have to targte both..
E**********e
发帖数: 1736
12
来自主题: DataSciences版 - 有关neuralnetwork基本的问题请教
在学neural network,简单的forward-feed 两层结构。 通晓了基本概念,像error
back propogation。 现在通过一个网上找来的列子在学习https://github.com/
dennybritz/nn-from-scratch,加深理解。但是run code的低时候,对输出层的error
的计算有点问题。
这个例子是个binary class 问题。 每个observation的target 是 0/1。当然这种类型
的问题可以用一个输出,然后预测的target就用一个logistic sigmoid来计算,然后算
cross-entropy error function function。 但是在这个例子用的是softmax 来计算,
也就是有输出有两个。 从开始到计算输出的两个值都好理解,也明白里面的algorithm
. 就是在计算输出值很target的差或erro是有点问题。 比如最后五个obervations通过
softmax的计算得到的概率是。
y_hat=
[ 0.67664077, 0.32335923],
... 阅读全帖
(共0页)