h
HiveMind
v4850ccalibraryNo description provided.
0 stars0 forksUnknownupdated 11 months ago
- #simd arrays in slices
- summing is difficult
- Would have to find an easy way to transpose to another SIMD equiv ideally without converting
- Multiple n,m builtin matrices
- keeping track and cumulnation of results is difficult
- I think this is last resort, could be very fast but the implementation is a nightmare
- can't go past 4x4 on anything OR
- A nightmare to do things past 4x4
- to do this you would make your first X number of calcs, then your next X number of calcs and add them all up at the end
- #soa or aos
- not sure yet haven't tried
- Simple Neural network implementation
- Generational learning?
- Society simulator
- Raylib
- Implement GUI visualizer
- Change to not use dynamic memory as much
- Implement a sigmoid variant
- How to calculate loss without base training data/ expected results
- Matrices
- Rows is each vector weight going to the neuron
- Col is the specific neuron
- rows must = length of input vector
- cols = number of neurons can be anything
- Odin is limited ot 4x4 builtin
Maintainer
