Hallo Shahzad,

please don't feel offended by me, but your circuit is a heavy misconstruction. I wonder why your friend was not reading datasheet.

Anyway, let's give help:

7812 regulator has a quiescent current of about 6mA. This current flows through ground terminal to 0V of your circuit. And although having a resistor here is sometimes used, in your case this resistor is much too big. 6mA accross 1kOhm gives a voltage drop of about 6V. So, your 12V regulator will NOT produce well regulated voltage. What you can see at output is something like input voltage minus 1V. 1V is that what drop across regulator.
Another fact is, that both regulators do not really see decoupling capacitors at there inputs AND outputs. If you 'isolate' them via diodes they do not work as decoupling capacitors. So, may be that Hans is right and your 5V regulator is oscillating?

I think it happens something like this: At input of 12V regulator there's a recitfied dc-voltage with some ripple. You will measure something arround 16V...18V, which is typical for 12V transformer used at current rating.The rest is total speculation: 12V regulator tries to produce 12V + (quiescent current x 1kOhm) which results in something undefined. May be it's something arround 12V, so you does not notice the design mistake? Possibly voltage at output of 12V regulator is much higher than 12V? Then again you have the problem with heat dissipation at 5V regulator.

The solution of cascaded regulating (first 12V regulator, then, from this 5V regulating) is widely used, when an additional transformer is to be saved. But this works only with very small supply currents! Something much less than 100mA, like Michael stated.

Bye,
Kai