Comments (12)
I suspect my Jaccard implementation is not liking all zero rows. That should not be too hard to fix, but I'll have to take a look and see what's actually going on there. Thanks for the report!
from umap.
That was my thought, initially -- I forgot to mention that I'm pretty certain I correctly did a test to clean out zero rows (not shown above)
Also added a column of 1s just as a sanity check while testing.
from umap.
Well I did certainly have a bug in handling of all zero rows, which is now fixed. It seems that isn't enough however, so now it's a matter of tracking down where the error actually is.
from umap.
Ah, it looks like it is in the random projection trees using cosine splitting. Let me look into that further ...
from umap.
It seems that the problem is when we try to split two identical points and the hyperplane is thus zero and we can't take the norm of it (although all zero vectors would also break this code, so I should fix that too). I will have to think a little about the "right" way to handle this.
from umap.
Ah, makes sense in light of my original data.
I'll try and familiarize myself more with umap.py in the meanwhile -- this is really interesting!
Looking forward to your paper.
from umap.
After some thought standard defensive programming will, in fact, do the right thing. I should have been more careful and caught these the first time, but hopefully this will work better now.
Edit: I tested this with your code above and it now works for me; I'll let you pull from master and reinstall and see it fixes the problems on your end as well.
from umap.
I suspect you intended for:
if right_norm == 0.0: right_norm = 0.0
to instead be:
if right_norm == 0.0: right_norm = 1.0
from umap.
Yes, sorry. I was working on a separate copy that "worked for me". All the more reason I appreciate the double check.
from umap.
No worries, happy to be able to help. Now to try the real data again!
from umap.
Not seeing any errors, closing ticket.
from umap.
Thank you!
from umap.
Related Issues (20)
- Implementation of sciki-learn's get_feature_names_out() API is not correct
- Is 'n_training_epochs' working for parameteric UMAP?
- visualize video data
- How to combine UMAP models in new data?
- Edit instructions to make them compatible with zsh
- Empty API page on UMAP API Guide? HOT 1
- PCA diagnostic error HOT 2
- Speed inquries HOT 2
- UMAP crashes when torch also imported before first run HOT 2
- Unable to pickle trained UMAP instance
- Reducing Model Size for UMAP on Large Datasets HOT 2
- umap.UMAP accepts strings as n_neighbors and min_dist, causing later failures
- Optimal dimensions
- RunUMAP Failing HOT 1
- Semi-deterministic output even though randon_state is set
- TypeError: Dispatcher._rebuild() got an unexpected keyword argument 'impl_kind' HOT 1
- illegal hardware instruction python HOT 2
- Transform new input with composite model HOT 1
- Inquiry on Utilizing UMAP for Text Similarity and Clustering HOT 4
- No clear documentation of default parameter values HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from umap.