AMD Instinct MI250X meets the minimum VRAM requirement for Q4 inference of dphn/dolphin-2.9.1-yi-1.5-34b. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
AMD Instinct MI250X can run dphn/dolphin-2.9.1-yi-1.5-34b with Q4 quantization. At approximately 161 tokens/second, you can expect Excellent speed - conversational response times under 1 second.
You have 111GB headroom, which is sufficient for system overhead and smooth operation.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 17GB | 128GB | 160.73 tok/s | ✅ Fits comfortably |
| Q8 | 35GB | 128GB | 107.15 tok/s | ✅ Fits comfortably |
| FP16 | 70GB | 128GB | 65.35 tok/s | ✅ Fits comfortably |