347 questions
0
votes
0
answers
93
views
CUDA error 700: an illegal memory access was encountered
I encounter error:
Application terminated with error: ??+0 (0x709D9F003D8A)
??+0 (0x709D9E884BA4)
??+0 (0x709D9E9F388C)
??+0 (0x709D9FF2FCF5)
??+0 (0x709D9FF31448)
??+0 (0x709D9EB84E21)
??+0 (...
- 1
1
vote
0
answers
82
views
How do I get predictions from my CatBoost models on Raspberry Pi using Python?
I've trained a couple of CatBoost models on my windows pc for both classification and regression targets. However, I'm having trouble installing the necessary dependencies on my Raspberry Pi 5 to ...
- 11
0
votes
1
answer
108
views
How do I increment a variable after each training iteration using the CatBoost Classifier in Python?
As the title says, I'm trying to increment a variable after each training iteration using the Catboost Classifier, to update a progress bar in a gui, and I can't seem to find anything about it on the ...
3
votes
1
answer
93
views
Concatenating TF-IDF Data and Categorical Data for CatBoost Model
I've been trying to concatenate TF-IDF data with categorical data. However, when concatenating, the categorical data is automatically converted to float by default. Since CatBoost doesn't support ...
0
votes
1
answer
166
views
Catboost error when loading pool from disk
I am creating a catboost pool from a pandas dataframe (columns have strings as names, not sure if thats relevant) and then quantizing it and saving to disk using this code:
import catboost as cb
...
- 363
0
votes
0
answers
164
views
XGBoost and LGBM models size depends on training data size for a given set of params whereas Catboost doesnt
I am comparing models in a walk forward cross validation setup, under python 3.11. For a given set of hyperparameters, xgboost and LGBM models size (when pickled or saved using the library saving ...
- 11
0
votes
2
answers
314
views
Using Optuna for CatBoost with batches: got nan on second trial
I am trying to tune CatBoost's hyperparameters using Optuna. I need to train my CatBoost model using batches, because training data is too big.
Here is my code:
def expand_embeddings(df, embedding_col=...
- 227
0
votes
1
answer
286
views
More efficient way to stream data to AWS Batch Transform Job
I have a sagemaker process for training and running inference on data in sagemaker:
processing job: read input csv files from s3 and clean up the data, output csv files to s3
processing job: read in ...
- 69
0
votes
1
answer
162
views
CatBoost crashes when launching Optuna
I want to tune a regressor catboost using Optuna and local GPU. The dataset is not very large: the training sample contains about 120k records and only 16 features (including categorical ones). I run ...
1
vote
0
answers
82
views
Unable to initialize Spark CatBoostClassifier with parameters
I am trying to create a CatBoostClassifier using catboost_spark. In the regular Catboost package, parameters such as learning_rate, loss_function, num_leaves, etc. can be included when creating the ...
- 11
1
vote
1
answer
643
views
Problem with importing `catboost` package
I was trying to install catboost, and everything was going well until I decided to upgrade to Python 3.12. After the upgrade, I encountered an error when I tried to import it:
numpy.dtype size changed,...
0
votes
1
answer
381
views
Catboost Feature Importance Calculation
I fitted a simple binary classification model with only 3 trees and wanted to check if feature importance results are similar to the formula in Catboost documentation (PredictionValuesChange).
After ...
- 86
1
vote
0
answers
151
views
Why does a trained CatBoost model need to provide training Pool if there are categorical features?
For many methods in CatBoost, for example:
plot_tree(tree_idx, pool=None)has an optional argument of pool. This argument becomes required if there are any one-hot encoded categorical variables.
...
- 11
2
votes
2
answers
6k
views
numpy.dtype size changed, may indicate binary incompatibility
I am using Numpy version 2.x. I get this error:
from catboost import CatBoostClassifier
File "_catboost.pyx", line 1, in init _catboost
ValueError: numpy.dtype size changed, may indicate ...
- 21
0
votes
0
answers
356
views
Catboost Installation using cargo rust
I am trying to follow this tutorial which helps serve a catboost model using rust. The installation of the catboost package is proving to be a challenge here.
I have tried doing
cargo install --git &...
- 375