Views
No views yet
Benchmark validation card — This repository documents the inference performance of mistralai/Mistral-7B-v0.3 when optimized with the RunuX-AI runtime on Google TPU v5e.
| Metric | PyTorch (torch_xla) | RunuX-AI | Improvement |
|---|---|---|---|
| Throughput | 21.5 tok/s | 67.1 tok/s | 3.12× faster |
| Energy | 9.30 J/tok | 2.98 J/tok | 3.12× lower |
| MXU Util | ~32% | 88% | 2.75× higher |
do_sample=False)1# Install baseline framework
2pip install torch torch_xla[tpu] transformers accelerate
3
4# Run baseline benchmark
5python benchmark_baselines.py --models mistral-7b-v0.31@article{callens2026runux,
2 title={RunuX-AI: Memory-Efficient, Energy-Aware Inference Runtime for Edge and Cloud Accelerators},
3 author={Callens, Xavier},
4 year={2026},
5 note={Socrate AI Lab}
6}