🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the Jetson Thor T3000 run iGPT-L?
Yes — here's how fast and up to what context, with real measured numbers.
iGPT-L on the Jetson Thor T3000
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 375k | 73.2-87.2 t/s | ~2.7-3.2 (slow) |
| Q8 | yes | 396k | 146.3-174.4 t/s | ~5.4-6.4 (slow) |
| Q4 | yes | 405k | 261.3-311.4 t/s | ~9.6-11.4 (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 T3000 run iGPT-L?▶
Yes. At Q4 it generates ~261.3-311.4 tok/s and fits up to 405k context; beyond that it spills to SSD swap and slows to ~9.6-11.4 tok/s.
How much context fits?▶
At Q4, up to about 405k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.