In this research, we propose a new low-precision framework, TENT, to leverage the benefits of a tapered fixed-point numerical format in TinyML models.
Additionally, the framework is amenable for different quantization approaches and supports mixed-precision floating point and fixed-point numerical formats.
Recently, the posit numerical format has shown promise for DNN data representation and compute with ultra-low precision ([5.. 8]-bit).
Our results indicate that posits are a natural fit for DNN inference, outperforming at $\leq$8-bit precision, and can be realized with competitive resource requirements relative to those of floating point.
We propose a precision-adaptable FPGA soft core for exact multiply-and-accumulate for uniform comparison across three numerical formats, fixed, floating-point and posit.
Conventional reduced-precision numerical formats, such as fixed-point and floating point, cannot accurately represent deep neural network parameters with a nonlinear distribution and small dynamic range.