LM317 Output Only 17.4mW to 2W MP3 Decoder—How to Increase Power Delivery?

Hi! I have this simple circuit that activates a load, whenever the switch is pressed (shorted). When the switch is released from being pressed, the load will power down slowly. I have already put this up on a breadboard. When the load is an LED, it works! But the actual load that I will use is a 2W MP3 Decoder. The load only receives 17.4 mW (based on my schematic diagram below, also see attachments) but before reaching the LM317, the power is 4.11 W. How can I make the load receive at least 2W. Please help! Thank you!

Schematic Diagram: http://i64.tinypic.com/2s0n71g.jpg

Your LM317 is set up to give about 1.75 V out.  The LM317 can supply 1A.  So theoretically your output power is 1.75V x 1A = 1.75W4W at 1.75V is over 2A.  If the circuit worked, this current would pass through a BC558 transistor which can take nowhere near this figure.  So if you did get the circuit working, you'd burn the BC558 out in short order....To get 4.17 V out, R5 would need to be 1.3K (1300 ohms).  I would short out Q4 and get this part of the circuit working properly before you go further. You should be able to get 2W out of it (0.5A @4V).Then use a better transistor for Q4, a TIP31 or similar should do.This is a very wasteful circuit, your LM317 is going to be dissipating 3 W or so and will need a good heatsink.

I am surprised the circuit gives you any voltage at all. When you turn on the switch you are applying 9V across the base of Q2. This should blow the transistor. I don't understand the configuration of Q4. If anything, the emitter should be connected to the 9V. In configuration shown or my suggestion, should Q2 successfully turn on you would have 9V across the CB or EB junction of Q4, which would again be catastrophic.
It looks to me that you are trying to introduce a time constant with R6 and C3. This may work but you need to introduce some current limiting resistors or your transistors will be destroyed. Then pay attention to David's comments.