New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the NVIDIA H200 run (ensemble): AWD-LSTM-DOC (fin) × 5 (WT2)?
Yes — here's how fast and up to what context, with real measured numbers.
(ensemble): AWD-LSTM-DOC (fin) × 5 (WT2) on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1951k | 9470.3-11286.5 t/s | ~78.9-94.1 (slow) |
| Q8 | yes | 1954k | 18940.5-22573 t/s | ~157.8-188.1 (slow) |
| Q4 | yes | 1955k | 33822.4-40308.9 t/s | ~281.9-335.9 (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 H200 run (ensemble): AWD-LSTM-DOC (fin) × 5 (WT2)?▶
Yes. At Q4 it generates ~33822.4-40308.9 tok/s and fits up to 1955k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~281.9-335.9 tok/s.
How much context fits?▶
At Q4, up to about 1955k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.