Rename experimental_analyze_model_and_dataset to analyze_model_and_dataset
1.4.0 - 2023-03-20#
Speed-up the computation of PDP and CEP in the model analysis tool.
Add compilation of model into .h file.
[JS port] Add “prefix” argument to model loading method.
Rename logging function from LOG to YDF_LOG to limit risk of collision with TF or Absl.
[JS port] Fix memory leak. Release emscripten objects.
1.3.0 - 2023-01-24#
Setting the generic hyper-parameter “subsample” is enough enable random subsampling (to need to also set “sampling_method=RANDOM”).
Improve the display of decision tree structures.
The Hyper-parameter optimizer field “predefined_search_space” automatically configures the set of hyper-parameters to explore during automatic hyper-parameter tuning.
Replaces the MEAN_MIN_DEPTH variable importance with INV_MEAN_MIN_DEPTH.
1.2.0 - 2022-11-18#
YDF can load TF-DF models directly (i.e. a TF model with a YDF model in the “assets” sub directory).
Expose confusion tables in a GBT model’s analysis.
Add the “compute_variable_importances” tool to compute variable importances on an already trained model.
Add the “experimental_analyze_model_and_dataset” tool to understand/analyze models.
1.1.0 - 2022-10-21#
Early stopping is no longer triggered during first iterations. The initial iteration for early stopping can be controlled with the new parameter
Benchmark inference tool does not require for the dataset to contain the label column.
The user can instruct the tokenizer to perform no tokenization at all.
Fix GRPC dependency to version 1.50.0.
The new documentation is live at ydf.readthedocs.io.
1.0.0 - 2022-09-07#
Go (GoLang) inference API (Beta): simple engine written in Go to do inference on YDF and TF-DF models.
Creation of html evaluation report with plots (e.g., ROC, PR-ROC).
Add support for Random Forest, CART, regressive GBT and Ranking GBT models in the Go API.
Add customization of the number of IO threads in the deployment proto.
0.2.5 - 2022-06-15#
Multithreading of the oblique splitter for gradient boosted tree models.
Support for pure serving model i.e. model containing only serving data.
Add “edit_model” cli tool.
Remove bias toward low outcome in uplift modeling.
0.2.4 - 2022-05-17#
Discard hessian splits with score lower than the parents. This change has little effect on the model quality, but it can reduce its size.
Add internal flag
hessian_split_score_subtract_parentto subtract the parent score in the computation of an hessian split score.
Add the hyper-parameter optimizer as one of the meta-learner.
The Random Forest and CART learners support the
0.2.3 - 2021-01-27#
Honest Random Forests (also work with Gradient Boosted Tree and CART).
Can train Random Forests with example sampling without replacement.
Add support for Focal Loss in Gradient Boosted Tree learner.
Incorrect default evaluation of categorical split with uplift tasks. This was making uplift models with missing categorical values perform worst, and made the inference of uplift model possibly slower.
0.2.2 - 2021-12-13#
The CART learner exports the number of pruned nodes in the output model meta-data. Note: The CART learner outputs a Random Forest model with a single tree.
The Random Forest and CART learners support the
SetLoggingLevelto control the amount of logging.
Fix tree pruning in the CART learner for regressive tasks.
0.2.1 - 2021-11-05#
Add example of distributed training at examples/distributed_training.sh
Use the median bucket split value strategy in the discretized numerical splitters (local and distributed).
Register the GRPC distribution strategy in :train.
0.2.0 - 2021-10-29#
Distributed training of Gradient Boosted Decision Trees.
maximum_model_size_in_memory_in_byteshyper-parameter to limit the size of the model in memory.
Fix invalid splitting of pre-sorted numerical features (make use to use midpoint).
0.1.5 - 2021-08-11#
Fix incorrect handling of CART pruning when validation set is empty. Previously, the whole tree would be erroneously pruned. Now, pruning is disabled if the validation set is not specified.
0.1.4 - ???#
Add training interruption in the abstract learner API.
Reduce the memory usage of the pre-sorted feature index.
Multi-threaded computation of the pre-sorted feature index.
Disable GBT’s early stopping if the validation dataset ratio is zero.
Pre-computed and cache the structural variable importances.
0.1.3 - 2021-05-19#
Register new inference engines.
0.1.2 - 2021-05-18#
Inference engines: QuickScorer Extended and Pred
0.1.1 - 2021-05-17#
Migration to TensorFlow 2.5.0.
0.1.0 - 2021-05-11#
Initial release of Yggdrasil Decision Forests.
CLI: train show_model show_dataspec predict infer_dataspec evaluate convert_dataset benchmark_inference utils/synthetic_dataset)
Learners: Gradient Boosted Trees (and derivatives), Random Forest (and derivatives), Cart.