Support for fixed point neuron. All parameters
(weights, filter parameters, etc) should be able to
have different fixed point pression. This way we can
look at the impact hardware implementations would have
on the network.
we've been discussing that several times. The main advantage
of that is when we have a hardware implementation (say,
FPGA) of neurons and want to be able to run the same network
on PC hardware.
Do you have FPGA hardware?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Logged In: YES
user_id=271199
we've been discussing that several times. The main advantage
of that is when we have a hardware implementation (say,
FPGA) of neurons and want to be able to run the same network
on PC hardware.
Do you have FPGA hardware?
Logged In: YES
user_id=617920
> Do you have FPGA hardware?
Yep, we have boards with
- virtex 2 1000
- virtex 2 3000
- spartan 2e 200
- atmel fpslic
- altera stratix 25