Search papers, labs, and topics across Lattice.
1
0
2
Fine-tuning vision-language models on a richly described dataset boosts their accuracy in interpreting subtle driver actions by 10 points, revealing a critical gap in existing training methods.
