-
I noticed that a new projector head MLP is added after loading the pre-trained MoCo v3 model. However, the parameters of this newly added component are also set to requires_grad=False.
My question …
-
As titled, by using `pip install -e .` at the root dir, we will automatically install `torch==2.3.0`.
And `fast_jl`, which is the library used for fast cuda random projector, will fail under this new…
-
Hi,
[angular_random_projection_split](https://github.com/lmcinnes/umap/blob/d286224a86dfe730c76f082c6f523b69b57bd0f8/umap/rp_tree.py#L29-L53), [euclidean_random_projection_split](https://github.com…
-
(Plotting offline)
Setting the projection to orthographic seems to prevent scene_camera settings being applied. This creates difficulty in saving as non-png images that aren't the default view.
``…
-
The density map generated with skymapper does not seem to follow the resolution set by the `nside` parameter with `healpy`.
I run the following code with `nside=32` (resolution of 110 arcmin) and `…
-
Random projection currently stores internal parameters as an internal object that is created in the cython layer but opaque to the python layer. This means pickling the random projection model will no…
-
# Description:
Hello! I appreciate the excellent work on benchmarking Performer and Longformer against the base Transformer. I’d like to propose the implementation of additional efficient Transformer…
-
Hi,
We noticed that the graph for random projection execution times is very interesting.
For some reason, when n=1553, the execution is orders of magnitude slower than when n=1550. i.e in our cas…
-
When plotting data using pcolormesh on a basemap projection (or a cartopy projection) I notice strange lines appear when I set the alpha value to less than 1.
Example code:
```
from mpl_toolkit…
-
# Summary
https://arxiv.org/abs/2211.15118 proposes a faster K-means++ algorithm via random projection to a subspace of smaller dimension. Should we consider integrating the algorithm here to Faiss…