3 ms·
> When developing ML models, you rarely train "just one". Depends on your field. In Reinforcement Learning you often really do train just one, at least on the
by nightcracker 6y ago
> When developing ML models, you rarely train "just one".
Depends on your field. In Reinforcement Learning you often really do train just one, at least on the same data set (since the data set often is dynamically generated based on the behavior of the previous iteration of the model).
- volta87 6y agoEven in reinforcement learning you can train multiple model with different data-sets concurrently and combine them for the next iteration.