New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the FuriosaAI RNGD run Selfish-RNN (SNT-ASGD) Stacked LSTMs?
Yes — here's how fast and up to what context, with real measured numbers.
Selfish-RNN (SNT-ASGD) Stacked LSTMs on the FuriosaAI RNGD
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 650k | 19345.2-23511.9 t/s | ~515.9-627 (slow) |
| Q8 | yes | 651k | 38690.5-47023.8 t/s | ~1031.7-1254 (slow) |
| Q4 | yes | 651k | 69090.1-83971.1 t/s | ~1842.4-2239.2 (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 FuriosaAI RNGD run Selfish-RNN (SNT-ASGD) Stacked LSTMs?▶
Yes. At Q4 it generates ~69090.1-83971.1 tok/s and fits up to 651k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1842.4-2239.2 tok/s.
How much context fits?▶
At Q4, up to about 651k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.