The goal is to create a machine that will allow quickly repeating
strikes from a hammer that can be adjusted to use 25, 50 and 100 lb
hammer strikes. The hammer should be able to strike at a minimum, 15 -
20 times per minute, though faster is better.
That's not particularly heavy, or exceptionally fast. As noted in ShadoCat's answer, such machines have been around for a very long time; the Wikipedia article on Trip Hammers describes significantly heavier hammers, and the YouTube video he linked demonstrates a quite sufficient strike rate.
So, if you really want to justify the name super heavy smasher, we'll need to think about how to turn this up to 11, in a way that could be done by medieval craftsmen if they only had the idea--i.e., no new materials, no new construction methods, just mechanisms they could already know about, put together in a novel way.
For a given amount of power, you can use wooden cogwheels or leather belts and pulleys to adjust gear ratios to either lift a light hammer quickly, for lots of strikes per minute, or lift a heavier hammer more slowly. If we want to lift heavier hammers, lift them higher, lift them more frequently, or just lift more hammers (so more smiths can work simultaneously), we need more power. So, how do we get the most power possible?
If you want to be able to operate on a regular schedule, you certainly won't want to use inconsistent wind power as your primary source. If the geography is right, however, you may be able to take advantage of wind as an auxiliary source (described in more detail below).
The simplest arrangement, which you could turn on and off at will, is to use all of that unskilled labor to run treadwheels to power the hammers. More treadwheels with more laborers running in them gives you more power for more, and heavier, hammers.
However, you can also multiply the water power available with the use of a dam. If the stream isn't big enough to run all the hammers, or the size of hammer, that you want just by sticking a water-wheel in it, you can construct a dam to store up the stream's energy, and then release it on-demand. Over a long period of time, your average power use will be restricted by the stream's natural flow rate, but if you aren't working 24/7, a dam will allow you to capture the energy of the falling water during off-periods (nights and weekends, so to speak), and add it to what's available during the workday. A system of sluice gates could also be used to divert water to multiple wheels, so as to control multiple hammers independently without complicated gearing or clutch systems, and to control the flow of water to adjust the wheel speed.
Now, if you have a convenient pond or lake below the hammermill, you can augment your water-power system with wind power, via pumped water storage. A flat lake is no good for water wheels, but it serves a recyclable water source for increasing the power of the stream; a windmill, located by the lake or up by the dam, can be used to pump water from the lake up into the dam, via an Archimedes screw or other pumping mechanism, thus converting inconsistent wind power into controllable, stored hydropower.
If the mountain stream is particularly long, it may be desirable to build multiple water-wheel powered hammermills in stages going down the slope, and possibly multiple control dams as well. Building multiple small dams would also be simpler than building one huge one--and if one fails, it doesn't shut down your whole industry. Cascading millworks of this sort, while not common, have been around since antiquity.
Including a die is no problem. In a pinch, you can make them out of cast iron in a sand or ceramic mold. That won't produce the hardest or most durable dies, and they may require periodic replacement, but it'd work. Just don't drop a cast iron die onto a cold workpiece by accident.
For smaller hammers (such as specified in the question), it would also be a relatively simple matter to swap out heads of various weights on a single hammer, possibly held in place with a peg system. These could be lifted with a tilt-hammer (or tail-helve, as described in the Wikipedia article) cam-and-lever system (cam pushing down on the opposite side of a lever from the hammer head). For heavier hammers (e.g., quarter-tonners), you probably wouldn't want to go to the trouble of trying to swap them out on a regular basis, but that shouldn't be a problem since the bigger hammers are a bonus feature anyway. Those heavier hammers would use a nose-helve cam-and-lever system (a cam lifting the end of the hammer shaft beyond the head, on the same side as the fulcrum) to get better leverage with less strain on the wooden beams. Again as described in the Wikipedia article, belly-helve hammers (with the cams acting between the pivot and hammer head) could be used for intermediate-weight hammers, but they don't need to be, and if you have any super-heavy smashers, it makes sense to minimize the number of different kinds of mechanisms that need to be built, and just stick with the nose-helve design for anything too heavy to use a tail-helve.