Details
Presenter(s)
Display Name
Qiwen Wang
- Affiliation
-
AffiliationUniversity of Michigan
- Country
-
CountryUnited States
Abstract
We study factors that could degrade model performance in analog RRAM in-memory-computing (IMC) systems, including limited array size, ADC resolution, on/off ratio, and device conductance variations. Different levels of architecture-aware training methods were developed to mitigate these factors and allow the system to achieve accuracy close to floating-point baseline with realistic device parameters.