Extreme Multi-label Classification on Eurlex-4K (Inference Time)
12.7Inference Latency (ms)AttentionXML-1
Evaluation Results
| Method | Links | |
|---|---|---|
| AttentionXML-1GPU=Nvidia V100, batch size=12021.10 | 12.7 | |
| XR-Transformer-1GPU=Nvidia V100, batch size=12021.10 | 22.3 | |
| LightXML-1GPU=Nvidia V100, batch size=12021.10 | 24.7 | |
| X-Transformer-1GPU=Nvidia V100, batch size=12021.10 | 48.2 |