We analyse the convergence of one-hidden-layer ReLU networks trained by gradient flow on n data points, when the input lies in very high dimension.

We prove quantitative convergence estimates for single layer neural networks in the NTK regime to gaussian processes at positive training time