// RUN โ€” FITS

Can the NVIDIA L40S 48GB run VILA1.5-13B?

Yes โ€” here's how fast and up to what context, with real measured numbers.

VILA1.5-13B on the NVIDIA L40S 48GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 96k 23.4-27.9 t/s ~1.1-1.3 (slow)
Q8 yes 178k 46.7-55.7 t/s ~2.2-2.6 (slow)
Q4 yes 214k 83.5-99.5 t/s ~3.9-4.6 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to system RAM offload (depends on your system RAM) and speed drops to the "beyond" figure.

FAQ

Can the NVIDIA L40S 48GB run VILA1.5-13B?โ–ถ

Yes. At Q4 it generates ~83.5-99.5 tok/s and fits up to 214k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~3.9-4.6 tok/s.

How much context fits?โ–ถ

At Q4, up to about 214k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.