⚡ WebGPU Benchmark Results (55.96x speedup)
#15
by
Xenova
- opened
| Batch Size | WASM (ms) | WebGPU (ms) |
| 1 | 1094.80 | 62.50 |
| 2 | 2182.00 | 339.20 |
| 4 | 4334.60 | 305.10 |
| 8 | 8240.40 | 364.30 |
| 16 | 16116.20 | 416.20 |
| 32 | 31906.50 | 570.20 |
- Model: Xenova/all-MiniLM-L6-v2
- Quantized: false
- Sequence length: 512
- Browser: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36
- GPU: vendor=nvidia, architecture=turing, device=, description=
Opening up chrome dev tools while running the benchmark slows down the WASM version, making the speedup larger!
