// RUN โ€” FITS

Can the Jetson Thor T2000 run German ELECTRA Large?

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

German ELECTRA Large on the Jetson Thor T2000

QuantFits?Max contextSpeedBeyond context
FP16 yes 187k 149.3-177.9 t/s ~10.9-13 (slow)
Q8 yes 192k 298.5-355.8 t/s ~21.8-26 (slow)
Q4 yes 194k 533.1-635.3 t/s ~38.9-46.4 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.

FAQ

Can the Jetson Thor T2000 run German ELECTRA Large?โ–ถ

Yes. At Q4 it generates ~533.1-635.3 tok/s and fits up to 194k context; beyond that it spills to SSD swap and slows to ~38.9-46.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 194k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.