Closed nihargupte-ph closed 11 months ago
For the most part we use the 2D flow for density estimation. However, sometimes it is useful to use the 15D flow either for debugging (or temporaily) for 1 detector networks. However, we can't specify the batch_size when sampling. This can be fixed by editing https://github.com/dingo-gw/dingo/blob/main/dingo/gw/importance_sampling/importance_weights.py#L121 to provide settings["batch_size"].
Feel free to edit this script. It's somewhat deprecated compared to dingo_pipe. In addition, dingo_pipe should be updated to allow for 14D or 15D density recovery.
For the most part we use the 2D flow for density estimation. However, sometimes it is useful to use the 15D flow either for debugging (or temporaily) for 1 detector networks. However, we can't specify the batch_size when sampling. This can be fixed by editing https://github.com/dingo-gw/dingo/blob/main/dingo/gw/importance_sampling/importance_weights.py#L121 to provide settings["batch_size"].