Ophir Yoktan
Ophir Yoktan
Trying to use WeightedMinHashGenerator with sparse vectors fails, as the code assumes the vectors are simple dense numpy vectors. the use case is using LSH with TF\IDF for weighing the...
When opening a notebook using google colab (for example https://colab.research.google.com/github/PythonFreeCourse/Notebooks/blob/master/week1/1_Strings_and_Output.ipynb) paragraph direction is ignored. using: ``` Hebrew Text ``` does work
re: https://github.com/PythonFreeCourse/Notebooks/issues/37
https://github.com/TeamHG-Memex/eli5/blob/4cd66ac5e0ffde0c002d8f46dc64d7091f49445f/eli5/lightgbm.py#L252 pred_leafs contains num_trees * len(X) values tree_info contains num_trees values so the values are identical only when there's a single data row to explain
in this line: https://github.com/TeamHG-Memex/eli5/blob/054e65688bd0bb738440da7f2f509d01afce0246/eli5/sklearn/utils.py#L49 there's an attempt to unpack the output of predict proba, but (at least in lightgbm) the return value is a matrix, so unpacking raises an exception...
the other way around work (at least doesn't crash) fsspec version 0.6.1 tensorflow-io 0.10.0 ``` import fsspec import tensorflow_io ``` ``` [libprotobuf ERROR external/com_google_protobuf/src/google/protobuf/descriptor_database.cc:172] Symbol name "proto3.proto3" conflicts with the...
a common idiom in cli is to have a flag (usually --, but this is taken here) that all text afterwards is passed as is (mostly for passing to other...
also posted in stack overflow: http://stackoverflow.com/questions/34599726/strange-error-when-using-optunity-maximize-structured-with-discrete-parameters I'm trying to use optunity (version 1.1.1) for hyperparameter optimization When I'm trying to use categorical parameters in the optimization, I get a strange...
Instead of running a fixed number of iterations, allow stopping the optimization if no (significant) progress was done for some time