Benchmarks
Leaderboard for wireless foundation model evaluations. Each entry links to the dataset protocol (splits/metrics) and the finetuning recipe used to generate the score. Scores are reported as mean ± std across 3 seeds.
Use filters to select a release and narrow the table; click column headers to sort. Multi-task datasets list subtasks as rows with their own metrics. Click a score to open the detailed evaluation and click recipe to open the finetuning command.
WavesFM v1.0
Multimodal release with a shared ViT backbone for IQ, spectrograms, CSI, and CIR, trained with masked wireless modeling.
- Supports IQ and image-like modalities via modality-specific input embeddings.
- Evaluated on RFP, INTD/INTC, HAS, RFS, RML2022, RADCOM, 5G NR positioning, UWB indoor/industrial localization, and DeepMIMO LoS/NLoS + beam prediction.
- Finetuning regimes: linear probing, partial fine-tuning (FT2), and LoRA (R=32, alpha=64).
13 results
| 5G NR Positioning | CSI | 5G NR Positioning | Mean Localization Error (m) | |||
| CommRad RF | Spectrogram | RF Signal Classification | Mean Per-Class Accuracy | |||
| DeepMIMO Beam Prediction | CSI | Beam Prediction | Mean Per-Class Accuracy | |||
| DeepMIMO LOS/NLOS | CSI | LoS/NLoS Classification | Mean Per-Class Accuracy | |||
| EfficientFi | CSI | Human Activity Sensing | Mean Per-Class Accuracy | |||
| ICARUS | IQ | Interference Detection | Interference Detection Accuracy | |||
| Interference Classification | Interference Classification Accuracy | |||||
| POWDER | IQ | RF Fingerprinting | Mean Per-Class Accuracy | |||
| RADCOM | IQ | Radar Signal Classification | Radar Signal Classification Accuracy | |||
| Modulation Classification | Modulation Classification Accuracy | |||||
| RML2022 | IQ | Modulation Classification | Mean Per-Class Accuracy | |||
| UWB Indoor Positioning and Tracking | CIR | UWB Indoor Positioning | Mean Localization Error (m) | |||
| UWB Industrial Localization | CIR | UWB Industrial Localization | Mean Localization Error (m) |