Teaching a Small Model to Recognise a Sport From Moving Dots

说明与用途

Laya 在 400 条运动片段上训练后达到了 36% 的准确率，并且对快照顺序敏感（顺序打乱时下降 8%）。

问题

Can a small language model learn to identify sports from sequences of player coordinates represented as moving dots, and does it capture temporal order?

结果

Fine‑tuned Laya achieved 36% accuracy on the 400‑clip benchmark, outperforming the official recipe and matching frontier models. It showed a clear dependence on snapshot order, losing 8 points when snapshots were shuffled, similar to MiniRocket and Opus 5.5.

工作流

• Prepare 400 clips of 10 dots over 8 snapshots as text coordinates
• Split data into five folds grouped by game
• Fine‑tune Laya on each fold using 32 epochs, selecting the epoch with best held‑out validation
• Evaluate accuracy on test clips for ordered, shuffled, and single‑snapshot conditions
• Compare results to MiniRocket, DeepSets, and frontier language models

证据

Untouched, Laya gets 24% of the clips right, which is chance.

来源

With 32 epochs and early stopping, 0.36 in order against 0.27 shuffled and 0.27 with a single snapshot.

来源

The second row is the one I was after. Shuffling the snapshots costs Laya 8 points, of the same order as the 10 it cost Opus 5.5.

来源