🔌
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 NVIDIA B200 run XLM-RoBERTa?
Yes — here's how fast and up to what context, with real measured numbers.
XLM-RoBERTa on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2656k | 5309.1-6327.3 t/s | ~26.5-31.6 (slow) |
| Q8 | yes | 2664k | 10618.2-12654.5 t/s | ~53.1-63.3 (slow) |
| Q4 | yes | 2668k | 18961-22597.4 t/s | ~94.8-113 (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 B200 run XLM-RoBERTa?▶
Yes. At Q4 it generates ~18961-22597.4 tok/s and fits up to 2668k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~94.8-113 tok/s.
How much context fits?▶
At Q4, up to about 2668k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.