5.0.0
Fix comment
Add float dotproduct implementation for NEON
Add TFloat data type for neural network
bugfix of FMA port to FAST_FLOAT
Add missing check for __ARM_NEON
Implement DotProductSSE() for FAST_FLOAT
Prepare using float instead of double for LSTM calculations
bugfixing the AVX2 Extract8+16 codes
Disable HAVE_FRAMEWORK_ACCELERATE for compilers which fail to compile with it