Loading content...
Loading GPU benchmarks...
AMD Instinct MI300X meets the minimum VRAM requirement for Q4 inference of skt/kogpt2-base-v2. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 4GB | 192GB | 752.62 tok/s | ✅ Fits comfortably |
| Q8 | 7GB | 192GB | 522.60 tok/s | ✅ Fits comfortably |
| FP16 | 15GB | 192GB | 279.87 tok/s | ✅ Fits comfortably |