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 ULM-FiT?
Yes — here's how fast and up to what context, with real measured numbers.
ULM-FiT on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1943k | 3972.8-4734.7 t/s | ~33.1-39.5 (slow) |
| Q8 | yes | 1950k | 7945.6-9469.4 t/s | ~66.2-78.9 (slow) |
| Q4 | yes | 1953k | 14188.5-16909.6 t/s | ~118.2-140.9 (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 ULM-FiT?▶
Yes. At Q4 it generates ~14188.5-16909.6 tok/s and fits up to 1953k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~118.2-140.9 tok/s.
How much context fits?▶
At Q4, up to about 1953k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.