⚡ WebGPU Benchmark Results M1 MBA (16.38x speedup)

#104
by Reza2kn - opened
Batch SizeWASM (int8)WASM (fp16)WASM (fp32)WebGPU (fp16)WebGPU (fp32)
1382.00397.00382.0038.0047.00
2768.00800.00762.0081.00114.00
41533.001610.001541.00110.00175.00
83070.003260.003119.00243.00387.00
166204.006736.006413.00430.00710.00
3212648.0013763.0012849.00840.001592.00
  • Model: Xenova/all-MiniLM-L6-v2
  • Tests run: WASM (int8), WASM (fp16), WASM (fp32), WebGPU (fp16), WebGPU (fp32)
  • Sequence length: 512
  • Browser: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/129.0.0.0 Safari/537.36
  • GPU: vendor=apple, architecture=common-3, device=, description=
![Screenshot 2024-09-30 at 2.56.19 PM.png](https://cdn-uploads.huggingface.co/production/uploads/64c1c77c245c55a21c6f5a13/nXUflB-kW7xZ4fSuLJA9u.png)

Sign up or log in to comment