Comments (6)
I set 'bst:gamma': 0.0
in case that matters
from xgboost.
hmm.. can you post a script that can produces the error? I can't produce this error with my script.
#!/usr/bin/python
import sys
import sklearn.datasets
sys.path.append('../')
import xgboost as xgb
x, y = sklearn.datasets.make_hastie_10_2()
y[y==-1] = 0
dtrain = xgb.DMatrix(x, label=y)
param = {'bst:max_depth':2, 'bst:eta':1, 'bst:gamma':0, 'objective':'binary:logistic', 'silent':1 }
num_round = 10
evallist = [(dtrain,'train')]
bst = xgb.train( param, dtrain, num_round, evallist )
from xgboost.
this is strange: I fail to create an isolated script to reproduce it. Even stranger: the error only occurs if I run either Rgbm (via rpy2) or sklearn's GradientBoostingClassifier before... but in this case its 100% reproducible. BTW: it happens in bst.predict
not xgb.train
from xgboost.
i have some idea what happened.
This can due to the fact that we are using pointer to check if a matrix a cached. If a matrix get freed and another allocation gets the same address, we can run into trouble.
There is a update in dev branch that fixed this problem, but we would like to test before we release it
from xgboost.
checked with the dev branch - you fixed it!
from xgboost.
Thanks peter! Indeed this is a problem that we overlooked:)
from xgboost.
Related Issues (20)
- Model with External memory doesn't work with categorical features - with reproducible experiment (Python) HOT 1
- XGBoostError: [15:33:38] C:\buildkite-agent\builds\buildkite-windows-cpu-autoscaling-group-i-0b3782d1791676daf-1\xgboost\xgboost-ci-windows\src\data\data.cc:277: All feature_types must be one of {int, float, i, q, c}. HOT 1
- Set DMatrix feature_names when passing in a dask dataframe HOT 2
- Weighting isnt applied when using custom/sklearn eval_metric callback via sklearn API HOT 1
- multi-output regression - memory overload HOT 2
- xgboost too slow HOT 10
- Are There Ways to Identify Which Subsample of Data an XGBoost Tree Was Fitted On? HOT 2
- Proposal: Consistent package naming convention HOT 2
- Support cuPy array in custom objective with the sklearn interface.
- segfault when `X` is `None` in `.fit()`
- KeyError: "None of [Int64Index([1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n ...\n 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],\n dtype='int64', length=163)] are in the [columns]" HOT 2
- Mismatch of split_conditions loading via `json.load` and `load_model` HOT 4
- Custom objective for multi-output regression with multi_strategy=multi_output_tree HOT 2
- Allow users to set the logger level in XGBoost-PySpark HOT 4
- Is there a way to update tree node split thresholds using new data without changes tree structures ?
- No response using xgboost after uwsgi integration with flask HOT 17
- exact tree training does not seem to colsample bynode HOT 2
- Invalid classes inferred from unique values of `y`. HOT 3
- RROR ml.dmlc.xgboost4j.java.RabitTracker:[AutoML] Uncaught exception thrown by worker: java.lang.InterruptedException: null HOT 1
- Start from previous boosting rounds for training continuation.
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 xgboost.