🔌
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 GeForce RTX 2050 run CollabLLM?
Yes — here's how fast and up to what context, with real measured numbers.
CollabLLM on the NVIDIA GeForce RTX 2050
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | no | — | — | — |
| Q8 | no | — | — | — |
| Q4 | yes | 4k | 18.3-21.8 t/s | ~1.6-1.9 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the NVIDIA GeForce RTX 2050 run CollabLLM?▶
Yes. At Q4 it generates ~18.3-21.8 tok/s and fits up to 4k context; beyond that it spills to SSD swap and slows to ~1.6-1.9 tok/s.
How much context fits?▶
At Q4, up to about 4k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.