Jay Wang

I can reproduce this problem on macOS, ```julia julia> @time using HDF5 0.415766 seconds (227.81 k allocations: 12.983 MiB) julia> @time using JLD 9.374953 seconds (2.00 M allocations: 117.350 MiB,...

Having the same issue with `umap_learn==0.5.3`.

The logo looks so cool! I thought the left padding of the badge could be increased a little bit, so I played around with the badge svg using some other...

I guess this is because JupyterLab has many [`focus()` calls](https://github.com/jupyterlab/jupyterlab/search?q=node.focus%28%29) inside the iframe. Each call makes the iframe's parent to focus on the iframe element, so that FireFox scrolls to...

I believe `2*self.M` here means sampling all combinations with 1 feature and all combinations with `self.M - 1` features. These combinations have the largest kernel weights.

I am facing the same issue. For me, the ONNX model's performance is way worse than the original XGBoost model even if I am not using early stopping. Did you...

@PatTheMav yeah, how do you change the powerline to "use normal unicode glyphs"? Can I do configure it with MacVim?

@ZyX-I Thank you so much, it works perfectly! 👍🏼

@jdsimcoe Sure. I just did what ZyX-I says above. Create a `config.json` file in the powerline configure directory `~/.config/powerline/` (if you don't have it, just create one). Then add `{"ext":{"vim":{"top_theme":"unicode"}}}`...

I am wondering how to trigger off the auto-preview feature please?