AMD Instinct MI210 meets the minimum VRAM requirement for Q4 inference of trl-internal-testing/tiny-LlamaForCausalLM-3.2. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
AMD Instinct MI210 can run trl-internal-testing/tiny-LlamaForCausalLM-3.2 with Q4 quantization. At approximately 254 tokens/second, you can expect Excellent speed - conversational response times under 1 second.
You have 60GB headroom, which is sufficient for system overhead and smooth operation.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 4GB | 64GB | 253.94 tok/s | ✅ Fits comfortably |
| Q8 | 7GB | 64GB | 157.85 tok/s | ✅ Fits comfortably |
| FP16 | 15GB | 64GB | 93.99 tok/s | ✅ Fits comfortably |