1 paper · 1 filter
Devleena Das, Rajeev Patwari, Ashish Sirasao
Inference optimizations such as quantization, pruning, format and datatype conversion, model export, and serialization can lead to functional degradations in language model task pe…