Basis vectors tend to depend on initialized value

microsoft / LQ-Nets

LQ-Nets: Learned Quantization for Highly Accurate and Compact Deep Neural Networks

MIT License

239 stars 69 forks source link

Basis vectors tend to depend on initialized value #6

Closed avinabsaha closed 5 years ago

avinabsaha commented 5 years ago

Tried implementing this on Pytorch. The basis values tend to be fixed to the initialized value and do not update much. Also network accuracy increases to 66% around and saturates after that.