TY  - RPRT
TI  - Adaptive Inference Batching using Policy Gradients
AU  - Ruslan Sharifullin
PY  - 2026
UR  - https://arxiv.org/abs/2607.05272
ID  - 2607.05272
ER  -