Trajectory Generation on MobilitySyn
0DistanceTrajGenAgent
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| TrajGenAgentBackbone=Qwen2.5-32B-Instruct, Generation mode=zero-shot, Temperature=0.902026.06 | 0 | 0.0051 | 0.1308 | 0 | 0.0003 | 0.0003 | 0 | |
| TransformerTraining epochs=200, Optimizer=Adam, Learning rate=0.001, Embedding size=256, Architecture=decoder-only, Layers=4, Attention heads=42026.06 | 0.0085 | 0.376 | 0.151 | 0.281 | 0.0047 | 0.0065 | 0.0115 | |
| SeqGANEmbedding dimension=16, Hidden units=16, Pre-training epochs=40, Adversarial epochs=202026.06 | 0.0089 | 0.0738 | 0.1344 | 0.2437 | 0.0035 | 0.0062 | 0.0108 | |
| GRUTraining epochs=200, Optimizer=Adam, Learning rate=0.001, Embedding size=256, Layers=6, Hidden units=5122026.06 | 0.0116 | 0.1859 | 0.1747 | 0.3368 | 0.0082 | 0.0097 | 0.0132 | |
| LSTMTraining epochs=200, Optimizer=Adam, Learning rate=0.001, Embedding size=256, Layers=6, Hidden units=5122026.06 | 0.0131 | 0.2823 | 0.168 | 0.3046 | 0.0044 | 0.0078 | 0.0135 | |
| Geo-LlamaBackbone=Llama-2-7b-chat-hf, Fine-tuning=LoRA, Batch size=48, Learning rate=0.00001, LoRA r=16, Epochs=20, Temperature=1.22026.06 | 0.0268 | 0.5528 | 0.0241 | 0.1209 | 0.0005 | 0.0005 | 0.0078 | |
| Geo-CETRABeam size=10, Top k=3, Optimizer=Adam, Learning rate=0.012026.06 | 0.0276 | 0.5784 | 0.0319 | 0.1573 | 0.0006 | 0.0006 | 0.0083 |