From 7921988a875ebb2fffc46f8cce13f05b53b25527 Mon Sep 17 00:00:00 2001 From: sarahmish Date: Tue, 8 Sep 2026 18:59:39 -0400 Subject: [PATCH] add synthesizer cards --- .../run_benchmark/upload_benchmark_results.py | 12 +----------- sdgym/synthesizer_descriptions.yaml | 19 +++++++++++++++++++ 2 files changed, 20 insertions(+), 11 deletions(-) diff --git a/sdgym/run_benchmark/upload_benchmark_results.py b/sdgym/run_benchmark/upload_benchmark_results.py index 0cc17193..e01639b7 100644 --- a/sdgym/run_benchmark/upload_benchmark_results.py +++ b/sdgym/run_benchmark/upload_benchmark_results.py @@ -33,17 +33,7 @@ from sdgym.utils import _set_column_width LOGGER = logging.getLogger(__name__) -SYNTHESIZER_TO_GLOBAL_POSITION = { - 'CTGAN': 'middle right', - 'TVAE': 'middle left', - 'GaussianCopula': 'bottom center', - 'Uniform': 'top center', - 'Column': 'top center', - 'CopulaGAN': 'top center', - 'RealTabFormer': 'bottom center', - 'HSA': 'bottom center', - 'Independent': 'top center', -} + SDGYM_RUNS_FILENAME = 'SDGym_Runs.xlsx' MODEL_DETAILS_FILENAME = 'Model_Details.xlsx' DATASET_DETAILS_FILENAME = 'Dataset_Details.xlsx' diff --git a/sdgym/synthesizer_descriptions.yaml b/sdgym/synthesizer_descriptions.yaml index 22f07c75..4475a25e 100644 --- a/sdgym/synthesizer_descriptions.yaml +++ b/sdgym/synthesizer_descriptions.yaml @@ -134,3 +134,22 @@ XGCSynthesizer: to train a model and generate synthetic data similar to the GaussianCopulaSynthesizer. However, it contains some additional features for higher quality modeling. +TabDDOMSynthesizer: + library: sdgym + organization: Yandex + modality: single_table + type: Deep Learning + description: > + The TabDDPM model is a denoising diffusion probabilistic model for tabular datasets. + It uses the multinomial diffusion to model the categorical and binary features, and + the Gaussian diffusion to model the numerical ones. + +ClavaDDPMSynthesizer: + library: sdgym + organization: University of Waterloo, Vector Institute, Royal Bank of Canada + modality: multi_table + type: Deep Learning + description: > + The ClavaDDPM model is an extension of the TabDDPM model to handle multi-table datasets. + It is based on clustering labels as intermediaries to model relationships between tables, + specifically focusing on foreign key constraints. \ No newline at end of file