Learning Rate - What happens if your learning rate is too high?

Setting a learning rate that's excessively high can lead to instability in your model's convergence; the training loss might fluctuate erratically, or possibly get trapped at a suboptimal level, also known as a local minima.