Update QuickNet documentation.

[AdamHillier]

This is a draft as we need to take new benchmarks using the 0.6 release.

[AdamHillier]

Will resolve #662.

[simonmaurer]

can provide the latency evaluation for mac mini (Apple M1) that we could directly add here. so an additional commit on this PR for the ReadMe.md would not be needed.

[simonmaurer]

Here's the latency results for the Apple mac mini 2020 (M1) using the latest LCE benchmark v0.6 including this PR

The table below presents single-/multi-threaded performance of Larq Compute Engine v0.6 on
different versions of QuickNet (trained on ImageNet dataset, released on Larq Zoo)
on an Apple mac mini 2020 (M1):

Model	Top-1 Accuracy	mac mini M1, ms (1 thread)	mac mini M1, ms (1 thread w/ XNNPACK)
QuickNetSmall (.h5)	59.4 %	4.0	3.5
QuickNet (.h5)	63.3 %	5.8	5.4
QuickNetLarge (.h5)	66.9 %	9.9	9.5

Model	Top-1 Accuracy	mac mini M1, ms (4 threads)	mac mini M1, ms (4 threads w/ XNNPACK)
QuickNetSmall (.h5)	59.4 %	1.8	1.9
QuickNet (.h5)	63.3 %	2.5	2.6
QuickNetLarge (.h5)	66.9 %	3.9	4.1

[lgeiger]

How do you feel about removing the links to the *.h5 files completely from the README here? I think these were useful when some of the models weren't available on Larq Zoo yet, but I don't think these add any value currently.

[AdamHillier]

Nice, yeah I've removed the h5 links now. I don't have my RPi to hand so can't record new benchmarks atm, but will update those numbers by tomorrow.

[CNugteren]

I don't have my RPi to hand so can't record new benchmarks atm, but will update those numbers by tomorrow.

Either @AdamHillier lives on a planet with a very long day-night cycle or he forgot about this, so I took the liberty to update the numbers for the RPi4. I also clarified that we run the device at 1.5GHz (some run at 1.8GHz), updated the RPi4 spec page link in the docs, and updated the benchmarking dates.

I think this is now ready for review.