// RUN โ€” FITS

Can the Instinct MI440X run StyleGAN2?

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

StyleGAN2 on the Instinct MI440X

QuantFits?Max contextSpeedBeyond context
FP16 yes 6041k 205800-251533.3 t/s ~420-513.3 (slow)
Q8 yes 6041k 411600-503066.7 t/s ~840-1026.7 (slow)
Q4 yes 6041k 735000-898333.3 t/s ~1500-1833.3 (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 Instinct MI440X run StyleGAN2?โ–ถ

Yes. At Q4 it generates ~735000-898333.3 tok/s and fits up to 6041k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1500-1833.3 tok/s.

How much context fits?โ–ถ

At Q4, up to about 6041k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.