List your company

ResearchFormerly WizWand

Inference Efficiency

Benchmarks

TaskDatasetSOTA resultTrendResults

Inference Efficiency

Inference Efficiency

80.8Throughput (Hz)

4

Speech Restoration

Inference Efficiency

0.015RTF

4

Egocentric Human Mesh Recovery

Inference Efficiency NVIDIA RTX 3090

4.74GFLOPs

3

Instance Segmentation

Inference Efficiency NVIDIA A100

54Throughput (Img/s)

3