⚡ WebGPU Benchmark Results (101.56x speedup)
#57
by
Xenova
HF staff
- opened
| Batch Size | WASM (fp32) | WebGPU (fp16) |
| 1 | 1032.40 | 73.10 |
| 2 | 2038.10 | 155.50 |
| 4 | 4102.40 | 207.30 |
| 8 | 8287.50 | 331.00 |
| 16 | 16153.70 | 168.80 |
| 32 | 32132.30 | 564.30 |
| 64 | 63950.70 | 629.70 |
- Model: Xenova/all-MiniLM-L6-v2
- Tests run: WASM (fp32), WebGPU (fp16)
- 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=