⚡ WebGPU Benchmark Results (1.00x speedup) - Ubuntu WebGPU fp32 up to bs=128
#55
by
pcuenq
HF staff
- opened
| Batch Size | WebGPU (fp32) |
| 1 | 17.70 |
| 2 | 40.40 |
| 4 | 32.50 |
| 8 | 75.70 |
| 16 | 91.80 |
| 32 | 211.50 |
| 64 | 339.70 |
| 128 | 768.20 |
- Model: Xenova/all-MiniLM-L6-v2
- Tests run: WebGPU (fp32)
- Sequence length: 512
- Browser: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36
- GPU: vendor=nvidia, architecture=lovelace, device=, description=