Please post an issue in Microsoft/LightGBM repository for any LightGBM issues you encounter. For critical issues (crash, prediction error, nonsense outputs…), you may also ping a member of the core team according to the relevant area of expertise by mentioning them with the arobase (@) symbol:
- @guolinke Guolin Ke (C++ code / R-package / Python-package)
- @chivee Qiwei Ye (C++ code / Python-package)
- @Laurae2 Damien Soukhavong (R-package)
- @jameslamb James Lamb (R-package)
- @wxchan Wenxuan Chen (Python-package)
- @henry0312 Tsukasa Omoto (Python-package)
- @StrikerRUS Nikita Titov (Python-package)
- @huanzhang12 Huan Zhang (GPU support)
Please include as much of the following information as possible when submitting a critical issue:
- Is it reproducible on CLI (command line interface), R, and/or Python?
- Is it specific to a wrapper? (R or Python?)
- Is it specific to the compiler? (gcc or Clang version? MinGW or Visual Studio version?)
- Is it specific to your Operating System? (Windows? Linux? macOS?)
- Are you able to reproduce this issue with a simple case?
- Does the issue persist after removing all optimization flags and compiling LightGBM in debug mode?
When submitting issues, please keep in mind that this is largely a volunteer effort, and we may not be available 24/7 to provide support.
- Question 1: Where do I find more details about LightGBM parameters?
- Solution 1: Take a look at Parameters and the Laurae++/Parameters website.
- Question 2: On datasets with millions of features, training does not start (or starts after a very long time).
- Solution 2: Use a smaller value for
bin_construct_sample_cntand a larger value for
- Question 3: When running LightGBM on a large dataset, my computer runs out of RAM.
- Solution 3: Multiple solutions: set the
histogram_pool_sizeparameter to the MB you want to use for LightGBM (histogram_pool_size + dataset size = approximately RAM used), lower
- Question 4: I am using Windows. Should I use Visual Studio or MinGW for compiling LightGBM?
- Solution 4: Visual Studio performs best for LightGBM.
- Question 5: When using LightGBM GPU, I cannot reproduce results over several runs.
- Solution 5: This is normal and expected behaviour, but you may try to use
gpu_use_dp = truefor reproducibility (see Microsoft/LightGBM#560). You may also use the CPU version.
- Question 6: Bagging is not reproducible when changing the number of threads.
- Solution 6: LightGBM bagging is multithreaded, so its output depends on the number of threads used. There is no workaround currently.
- Question 7: I tried to use Random Forest mode, and LightGBM crashes!
- Solution 7: This is expected behaviour for arbitrary parameters. To enable Random Forest,
you must use
feature_fractiondifferent from 1, along with a
bagging_freq. This thread includes an example.
- Question 8: CPU usage is low (like 10%) in Windows when using LightGBM on very large datasets with many-core systems.
- Solution 8: Please use Visual Studio as it may be 10x faster than MinGW especially for very large trees.
Question 9: When I’m trying to specify a categorical column with the
categorical_featureparameter, I get the following sequence of warnings, but there are no negative values in the column.
[LightGBM] [Warning] Met negative value in categorical features, will convert it to NaN [LightGBM] [Warning] There are no meaningful features, as all feature values are constant.
Solution 9: The column you’re trying to pass via
categorical_featurelikely contains very large values. Categorical features in LightGBM are limited by int32 range, so you cannot pass values that are greater than
Int32.MaxValue(2147483647) as categorical features (see Microsoft/LightGBM#1359). You should convert them to integers ranging from zero to the number of categories first.
Question 10: LightGBM crashes randomly with the error like this.
OMP: Error #15: Initializing libiomp5.dylib, but found libomp.dylib already initialized. OMP: Hint: This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to ensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the OpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment variable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently produce incorrect results. For more information, please see http://www.intel.com/software/products/support/.
Solution 10: File extensions in the error message may differ depending on the operating system. This error means that you have multiple OpenMP libraries installed on your machine and they conflict with each other.
If you are using Python distributed by Conda, then it is highly likely that the error is caused by the
numpypackage from Conda which includes the
mklpackage which in turn conflicts with the system-wide library. In this case you can update the
numpypackage in Conda or replace the Conda’s OpenMP library instance with system-wide one by creating a symlink to it in Conda environment folder
Assuming you are using macOS with Homebrew, the command which overwrites OpenMP library files in the current active Conda environment with symlinks to the system-wide library ones installed by Homebrew:
ln -sf `ls -d "$(brew --cellar libomp)"/*/lib`/* $CONDA_PREFIX/lib
If this is not your case, then you should find conflicting OpenMP library installations on your own and leave only one of them.
Question 11: LightGBM hangs when multithreading (OpenMP) and using forking in Linux at the same time.
Solution 11: Use
nthreads=1to disable multithreading of LightGBM. There is a bug with OpenMP which hangs forked sessions with multithreading activated. A more expensive solution is to use new processes instead of using fork, however, keep in mind it is creating new processes where you have to copy memory and load libraries (example: if you want to fork 16 times your current process, then you will require to make 16 copies of your dataset in memory) (see Microsoft/LightGBM#1789).
An alternative, if multithreading is really necessary inside the forked sessions, would be to compile LightGBM with Intel toolchain. Intel compilers are unaffected by this bug.
For C/C++ users, any OpenMP feature cannot be used before the fork happens. If an OpenMP feature is used before the fork happens (ex: using OpenMP for forking), OpenMP will hang inside the forked sessions. Use new processes instead and copy memory as required by creating new processes instead of forking (or, use Intel compilers).
- Question 1: Any training command using LightGBM does not work after an error occurred during the training of a previous LightGBM model.
- Solution 1: Run
lgb.unloader(wipe = TRUE)in the R console, and recreate the LightGBM datasets (this will wipe all LightGBM-related variables). Due to the pointers, choosing to not wipe variables will not fix the error. This is a known issue: Microsoft/LightGBM#698.
- Question 2: I used
setinfo, tried to print my
lgb.Dataset, and now the R console froze!
- Solution 2: Avoid printing the
setinfo. This is a known bug: Microsoft/LightGBM#539.
Question 1: I see error messages like this when install from GitHub using
python setup.py install.
error: Error: setup script specifies an absolute path: /Users/Microsoft/LightGBM/python-package/lightgbm/../../lib_lightgbm.so setup() arguments must *always* be /-separated paths relative to the setup.py directory, *never* absolute paths.
Solution 1: This error should be solved in latest version. If you still meet this error, try to remove
lightgbm.egg-infofolder in your Python-package and reinstall, or check this thread on stackoverflow.
Question 2: I see error messages like
Cannot get/set label/weight/init_score/group/num_data/num_feature before construct dataset
but I’ve already constructed a dataset by some code like
train = lightgbm.Dataset(X_train, y_train)
or error messages like
Cannot set predictor/reference/categorical feature after freed raw data, set free_raw_data=False when construct Dataset to avoid this.
Solution 2: Because LightGBM constructs bin mappers to build trees, and train and valid Datasets within one Booster share the same bin mappers, categorical features and feature names etc., the Dataset objects are constructed when constructing a Booster. If you set
free_raw_data=True(default), the raw data (with Python data struct) will be freed. So, if you want to:
- get label (or weight/init_score/group) before constructing a dataset, it’s same as get
- set label (or weight/init_score/group) before constructing a dataset, it’s same as
- get num_data (or num_feature) before constructing a dataset, you can get data with
self.data. Then, if your data is
numpy.ndarray, use some code like
- set predictor (or reference/categorical feature) after constructing a dataset,
you should set
free_raw_data=Falseor init a Dataset object with the same raw data
- get label (or weight/init_score/group) before constructing a dataset, it’s same as get
Question 3: I encounter segmentation faults (segfaults) randomly after installing LightGBM from PyPI using
pip install lightgbm.
Solution 3: We are doing our best to provide universal wheels which have high running speed and are compatible with any hardware, OS, compiler, etc. at the same time. However, sometimes it’s just impossible to guarantee the possibility of usage of LightGBM in any specific environment (see Microsoft/LightGBM#1743).
Therefore, the first thing you should try in case of segfaults is compiling from the source using
pip install --no-binary :all: lightgbm. For the OS-specific prerequisites see this guide.
Also, feel free to post a new issue in our GitHub repository. We always look at each case individually and try to find a root cause.