lmcinnes / pynndescent

A Python nearest neighbor descent for approximate nearest neighbors

lmcinnes/pynndescent Issues

Very high memory usage
Updated 2 months ago4
pynndescent is changing the input data, is this expecte?
Updated 3 months ago7
Querying the training set: runtime tradeoff for large k
Updated 4 months ago2
How to navigate pyinstaller
Closed 4 months ago1
pynndescent might break with next numba release
Closed 4 months ago3
Reverse diversification is actually forward diversification again
Updated 5 months ago
TSSS missing a factor of 2
Updated 6 months ago
true_angular is not a distance?
Updated 7 months ago
Minor bias in split selection?
Updated 8 months ago1
API to save and load index from disk
Updated 9 months ago
[BUG] Error out when using 'cosine' distance metrics (ufunc 'correct_alternative_cosine' did not contain a loop with signature matching types)
Updated 10 months ago10
Slice error using mac M1-max ARM
Updated 10 months ago6
Importing pickled index gives error - 'NNDescent' object has no attribute 'shape'
Updated a year ago1
Specifying threshold in distance metrics
Updated a year ago
Exceedingly large amount of memory usage
Closed a year ago
Newest version breaks with UMAP
Updated a year ago3
Tests fail: E SystemError: initialization of _internal failed without raising an exception
Updated a year ago
Sample identifiers for semantic search
Updated a year ago2
Question about covariance matrix used when using Mahalanobis distance
Updated a year ago
uint8 as internal data
Updated a year ago1
Cosine metric - error "Negative values in data passed to precomputed distance matrix"
Updated a year ago2
access distances in heap
Updated a year ago3
`make_dense_tree()` with `angular=True` can segfault on poorly-behaved datasets
Closed a year ago1
TypeError: np.matrix is not supported with sklearn 1.2.0
Closed a year ago2
How to have a payload associated with each vector?
Closed 2 years ago
Fatal Python error: Segmentation faultFatal Python error: Segmentation fault
Updated 2 years ago1
self implemented distance metric
Updated 2 years ago3
Question about the "epsilon" parameter in the query function.
Closed 2 years ago2
Correlation function needs reconsideration
Closed 2 years ago6
When distance computation is expensive how to gradually build graph
Updated 2 years ago3
Can't pass parameters to weighted_minkowski distance metric
Updated 2 years ago3
Distributed PyNNDescent
Updated 2 years ago1
NNDescent.update_with_changed_data needs improvements
Closed 2 years ago4
nested loop sequential or parallelization runtime
Updated 2 years ago1
Cannot use random projection trees
Updated 2 years ago8
NNDescent.update(X) not working as expected
Closed 2 years ago7
Related project: numba-kdtree
Updated 2 years ago1
scipy 1.8.0 causes: ValueError: Unknown Distance Metric: wminkowski in test suite
Updated 2 years ago1
AttributeError: 'NNDescent' object has no attribute 'parallel_batch_queries'
Updated 2 years ago1
Is it possible to speed up index generation for time series data?
Updated 2 years ago1
NNDescent.update() fails even for compressed or uncompressed indexes
Updated 2 years ago1
RFC: Deprecate support for Python 3.6
Closed 2 years ago4
Question about simple RP-Tree experiment in document
Updated 2 years ago
Measuring minkowski with p= 0.3
Closed 2 years ago10
is `pruning_degree_multiplier` ever used?
Closed 2 years ago2
How to store graph for further use ?
Closed 2 years ago4
Scripts for making plots on pynndescent docs
Closed 2 years ago3
Import fails
Updated 3 years ago
Metric for number of common non-zeros in sparse matrix
Closed 3 years ago2
Nearest neighbors for a subset of points
Updated 3 years ago3