Re: [Jprogramming] A Neural Network in 11 lines of Python (Part 1)

2017-11-18 Thread 'Skip Cave' via Programming
Raul, Thanks, that's exactly what I needed. I was trying to figure out what syn0 = 2*np.random.random((3,1)) - 1 did. Now I know. Running the network 1 times: Raul: tiny_toy_network1 '' 0.00966589 0.00786586 0.993588 0.992118 Article: Output After Training: [[ 0.00966449] [ 0.007

Re: [Jprogramming] Partitions

2017-11-18 Thread Erling Hellenäs
Hi all! I wondered what would happen when we ran out of digits.    a=:14 parMD 15    1{a 01023456789:;<= Cheers, Erling Hellenäs On 2017-11-17 19:14, '[email protected]' via Programming wrote: Erling Helenas, Raul Miller, and others have come up with various methods to generate sub

Re: [Jprogramming] Partitions

2017-11-18 Thread Raul Miller
Yeah, that's ascii for you - though it doesn't really matter. Of course, if that bothers you you could replace a=. a. }.~ a. i. '0' NB. symbol list - arbitrary, but '012...' with a=. ~.;{:@;: ::a:"1 '0',.~a. But that's probably not worth the extra millisecond it costs. Thanks, -- Raul

Re: [Jprogramming] Partitions

2017-11-18 Thread 'Mike Day' via Programming
Yes, choice of symbols is a matter of taste or something... Mike Please reply to [email protected]. Sent from my iPad > On 18 Nov 2017, at 12:36, Raul Miller wrote: > > Yeah, that's ascii for you - though it doesn't really matter. > > Of course, if that bothers you you could re

Re: [Jprogramming] Partitions

2017-11-18 Thread 'Mike Day' via Programming
Had to be a better way, though its performance isn’t crucial, of course. Thanks, Mike Please reply to [email protected]. Sent from my iPad > On 18 Nov 2017, at 01:07, Raul Miller wrote: > > blockmake=: (-@[ {."1&.> ] <@({.,"0 1 }.)"0 1~ 1+])i. > (5 blockmake 4) -: 4 makeblo