Find research datasets worth reusing

Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.

251

datasets available to search

ShareScore release 0.7.1

Dataset results

251 results for “deep learning models”

Learn how ShareScore rates datasets ↗

zenodo36/100

Surface Ozone, NO2, and PM2.5 Concentrations Estimated by the Deep Learning model (Air Transformer) based on Satellite data.

Surface ozone, NO2, and PM2.5 concentrations Estimated by the deep learning model (Air Transformer) based on massive ground-level monitoring, satellite observations, meteorological conditions, dynamic industrial emissions, and other ancillary data from May 2018 to June 2021.

opencc-by-4.0Nov 2023View details →

zenodo36/100

dataset- Tuberculosis detection using Squid Game Optimization with Deep Learning Model on Chest X-Ray Images

Open the record for dataset details and reuse information.

opencc-by-4.0Dec 2023View details →

zenodo36/100

Data from: Genome-scale annotation of protein binding sites via language model and geometric deep learning

The dataset contains the training and test sets of protein binding sites with DNA, RNA, peptide, protein, ATP, HEM, Zn2+, Ca2+, Mg2+ and Mn2+. Each protein is associated with 3 lines indicating the protein name (PDB accession code and chain), sequence and residue labels (0 for non-binding and 1 for binding), respectively. The ESMFold-predicted structures are also provided.

openmit-licenseMar 2024View details →

zenodo36/100

"iDCNNPred: An interpretable deep learning model for virtual screening and identification of PI3Ka inhibitors against triple-negative breast cancer"

In this study, we proposed a novel interpretable deep convolutional neural network prediction (iDCNNPred) system for classifying molecular bioactivity and identifying predictive potential inhibitors for the PI3Ka isoform protein. This system utilizes 2D molecular image representation as input features, instead of traditional molecular fingerprints or descriptors. The datasets used for model construction, prediction and screening of chemical library are provided in this uploaded data in <a href="../api/records/10947610/draft/files/Molecular_image_Custom_DCNN_datasets.zip/content" target="_blank" rel="noopener noreferrer">Molecular_image_Custom_DCNN_datasets.zip</a> file for Custom-DCNN models and <a href="../api/records/10947610/draft/files/Molecular_image_pre_trained_datasets.zip/content" target="_blank" rel="noopener noreferrer">Molecular_image_pre_trained_datasets.zip</a> file for Pre-trained fine-tuned models. The final run of models results given in file <a href="../api/records/10947610/draft/files/Custom_DCNN_Pre_trained_models.zip/content" target="_blank" rel="noopener noreferrer">Custom_DCNN_Pre_trained_models.zip</a>

opencc-by-4.0Apr 2024View details →

zenodo36/100

Cryo-EM and X-ray crystallography ligands represented as 3D voxel grids for training deep learning models

Ligand datasets used to train and evaluate the models studied in "Ligand Identification using Deep Learning" by Karolczak, J. et al. The blobs_full.tar.gz and cryoem_blobs.zip files contain compressed 3D numpy arrays (*.npz) of all the ligand blobs extracted from X-ray and cryo-EM PDB deposits prior to quality filtering. The npz file names correspond to the PDB ID, chain, residue number, and ligand name of the extracted blob. The cmb_data.csv file contains the tabular data used to train the CheckMyBlob model. The X-ray data were later divided into training and testing subsets according to the xray_train.csv and xray_holdout.csv files, respectively. The ligand_mapping.csv file contains the mapping from ligand IDs to ligand group names. Finally, the cryoem_qscores.csv file contains Q-scores that were used to filter cryo-EM ligands.

opencc-by-4.0Apr 2024View details →

zenodo36/100

Optimizing Deep Learning Models for Aflatoxin Detection: A Case of Artificial Intelligence-Driven Classified Groundnut Image Datasets for Postharvest Management

DATASET DESCRIPTION  This dataset comprises a curated collection of classified groundnut images, specifically designed for deep learning applications in aflatoxin detection. The dataset is organized into four distinct categories: Healthy, Moldy, Insect-Infested, and Physiological Disorder, making it a vital resource for training AI and machine learning models aimed at advancing agricultural research. These classifications are crucial for the development of AI-driven solutions addressing aflatoxin contamination, enhancing crop quality assessments, and improving postharvest management practices. The dataset has been developed to support research in agricultural Artificial Intelligence (AI), machine learning (ML), and food safety, with a focus on aiding resource-constrained regions in combating postharvest losses due to contamination. By leveraging this dataset, researchers can contribute to safeguarding public health, promoting food security, and supporting smallholder farmers. POTENTIAL APPLICATIONS This dataset provides numerous opportunities for innovation in agriculture through AI and deep learning technologies. Its key applications include: Early Aflatoxin Detection: Facilitates the development of AI-powered models for prompt identification of aflatoxins in groundnuts, helping mitigate associated health risks. Postharvest Management Improvement: Enables the creation of innovative solutions to enhance storage, handling, and processing, reducing contamination and losses. Food Safety and Quality Assurance: Strengthens agricultural value chains by supporting the production of safe and high-quality food products. BROADER IMPACT This resource is invaluable for fostering AI innovation in agriculture, particularly in resource-limited environments. It addresses critical challenges such as postharvest losses and food contamination while contributing to global efforts in sustainable agricultural development. By utilizing this dataset, researchers can improve food security, support smallholder farmers, and drive advancements in agricultural practices that benefit both local and global communities.

opencc-by-4.0Nov 2024View details →

zenodo36/100

APNet, an explainable sparse deep learning model to discover differentially active drivers of severe COVID-19

Motivation: Computational analyses of plasma proteomics provide translational insights into complex diseases such as COVID-19 by revealing molecules, cellular phenotypes, and signaling patterns that contribute to unfavorable clinical outcomes. Current in silico approaches dovetail differential expression, biostatistics, and machine learning, but often overlook nonlinear proteomic dynamics, like post-translational modifications, and provide limited biological interpretability beyond feature ranking.Results: We introduce APNet, a novel computational pipeline that combines differential activity analysis based on SJARACNe co-expression networks with PASNet, a biologically-informed sparse deep learning model to perform explainable predictions for COVID-19 severity. Co-expression and classification weights are ingested by the APNet driver-pathway network to aid result interpretation and hypothesis generation. APNet outperforms alternative models in patient classification across three COVID-19 proteomic datasets, identifying predictive drivers and pathways, including some confirmed by single-cell omics and highlighting under-explored biomarker circuitries in COVID-19.Availability and Implementation: APNet's R, Python scripts and Cytoscape methodologies are available at <a href="https://github.com/BiodataAnalysisGroup/APNet">https://github.com/BiodataAnalysisGroup/APNet</a>

opencc-by-4.0Dec 2023View details →

zenodo36/100

Are deep learning models in hydrology entity aware?

This is the dataset and code accompanying the publication "Are deep learning models in hydrology entity aware?" of Benedikt Heudorfer, Hoshin V. Gupta, and Ralf Loritz to allow reproducible results.

opencc-by-4.0Oct 2024View details →

zenodo36/100

Data and models for "Center-fixing of tropical cyclones using uncertainty-aware deep learning applied to high-temporal-resolution geostationary satellite imagery" by Lagerquist et al.

The file geocenter_models.tar contains all models comprising the GeoCenter ensemble: 3 convolutional neural networks (CNN), 3 isotonic-regression files (one for correcting each CNN’s mean estimate), and 3 more isotonic-regression files (one for correcting each CNN’s ensemble spread). Every model is found in a subdirectory whose names indicate which infrared (IR) wavelengths are used as input to the CNN. For example: <ul> <li> wavelengths-microns=3.900-7.340-13.300/model.weights.h5: An HDF5 file containing the trained CNN that uses data from bands 7, 10, 16 (corresponding to 3.9, 7.34, and 13.3 microns on the GOES ABI imager). The trained CNN can always be read by neural_net_utils.read_model() in the ml4tccf library (https://doi.org/10.5281/zenodo.15116854). </li> <li> wavelengths-microns=3.900-7.340-13.300/model_metadata.p: A Pickle file containing metadata for the trained CNN. This file is needed to read the CNN itself with neural_net_utils.read_model(). Otherwise, you will probably never need to access this metafile directly. </li> <li> wavelengths-microns=3.900-7.340-13.300/isotonic_regression/isotonic_regression.dill: A Dill file containing isotonic-regression models used to bias-correct the ensemble mean from the same CNN. The trained isotonic-regression models can always be read by scalar_isotonic_regression.read_file() in the ml4tccf library. Note that there are technically two isotonic-regression models for every CNN’s ensemble mean: one that bias-corrects the x-coordinate of the TC-center, another that bias-corrects the y-coordinate. </li> <li> wavelengths-microns=3.900-7.340-13.300/uncertainty_calibration/uncertainty_calibration.dill: A Dill file containing isotonic-regression models used to bias-correct the ensemble spread from the same CNN. In the ml4tccf code, I make a distinction between “isotonic_regression” (correcting the ensemble mean) and “uncertainty_calibration” (correcting the ensemble spread), but note that both models are isotonic regression and use the sklearn.isotonic.IsotonicRegression class. The trained uncertainty-calibration models can always be read by scalar_uncertainty_calibration.read_file() in the ml4tccf library. Again, note that there are technically two uncertainty-calibration models per CNN: one for spread in the x-coordinate, one for spread in the y-coordinate. </li> </ul>   As mentioned above, every trained CNN can be read by neural_net_utils.read_model(). Also, every trained CNN can be applied to new data (inference mode) by neural_net_utils.apply_model(). The input argument model_object should be the object returned by neural_net_utils.read_model(), and I suggest setting num_examples_per_batch = 10 to avoid out-of-memory errors. The only other input argument is predictor_matrices, which is a list of two numpy arrays. The first numpy array contains IR imagery centered at the first-guess TC center, and the second numpy array contains ATCF scalars. The first numpy array should have dimensions S (number of TC samples) x 300 (grid rows) x 300 (grid columns) x 9 (lag times) x 3 (wavelengths). Lag times should be in the following order: 240, 210, 180, 150, 120, 90, 60, 30, 0 min ago.  Wavelengths should be in the order indicated by the subdirectory name.  The numpy array itself should contain normalized brightness temperatures at the given lag times and wavelengths, following the grid specifications laid out in the journal paper (a plate carrée grid with 2-km spacing). The original IR data (brightness temperatures) must be normalized to z-scores using the same normalization parameters as in the journal paper, i.e., those based on the training data. See details below. The second numpy array in predictor_matrices should have dimensions S (number of TC samples) x 9 (variables). The variables must in the order: absolute latitude, cosine of longitude, sine of longitude, TC intensity, minimum central pressure, tropical flag, subtropical flag, extratropical flag, disturbance flag. The journal paper contains details on all these variables in one table. These variables must come from A-deck files at the second-most recent synoptic time. Like the IR data, these ATCF scalars must be normalized to z-scores using the same normalization parameters as in the journal paper. See details below.   Once you have predictions (estimated TC-center locations) from a CNN, you can bias-correct these predictions. To read the isotonic-regression model for the given CNN’s ensemble mean, use scalar_isotonic_regression.read_file() in the ml4tccf library. To apply the same model, use scalar_isotonic_regression.apply_models(). For the CNN’s ensemble spread, use scalar_uncertainty_calibration.read_file() and scalar_uncertainty_calibration.apply_models().   To normalize the IR data, you will need the file ir_satellite_normalization_params.tar included with this dataset. Within the tar file is a single zarr file. You can read the zarr file with normalization.read_file() in the ml4tccf library; then you can normalize new data with normalization.normalize_data().   To normalize the ATCF data, you will need the file a_deck_normalization_params.nc included with this dataset. This is a NetCDF file, containing the full set of training values for all 5 ATCF variables that are normalized (the binary storm-type flags are not normalized). You can read this file using any of the standard Python methods for reading NetCDF files, such as xarray.open_dataset(). To normalize new ATCF data, you can use the method normalization._normalize_one_variable(), where the argument actual_values_training is the list of training values from a_deck_normalization_params.nc for the given variable, while actual_values_new is the list of values to be normalized (currently in physical units, to be converted to z-score units).

opencc-by-4.0Sep 2024View details →

zenodo36/100

molxspec: Deep learning models for predicting MS2 spectra from molecular structures

This repository contains a pre-processed dataset derived from the <a href="https://gnps.ucsd.edu">GNPS public repository</a> of natural product mass spectra as well as pretrained model weights for four different types of model architectures using pytorch (version 1.9.0). The contents are as follows: <ul> <li>gnps_processed_data.tgz: Contains tab separated files of molecule/MS2 spectra pairs derived from GNPS after filtering for invalid structures, too large molecules (bigger than 2000 M/Z spectra), and structures that yielded valid 3D geometry optimization. The processing steps were done for positive ionization mode (pos_* files), though negative ionization data is also included (neg_* files)</li> <li>models.tgz: Contains pytorch format pretrained models for four different architecutres: MLP (a residual block multilayer perceptron trained on ECFP molecular fingerprints), BERT (the same MLP but trained on pretrained representations from the Zinc V1 pretrained ChemBERTa models on SMILES), GCN (a graph convolution architecture), and EGNN (an equivariant graph neural network). Models were trained on pos_processed_gnps_shuffled_with_3d_train.tsv found in the gnps_processed_data.tgz file described previously.</li> </ul>

opencc-by-4.0Nov 2021View details →

zenodo36/100

PIGNet: A physics-informed deep learning model toward generalized drug-target interaction predictions

Training, test datasets of the paper "PIGNet: A physics-informed deep learning model toward generalized drug-target interaction predictions".

opencc-by-4.0Feb 2022View details →

zenodo36/100

CAST Strong Lensing Finder Deep Learning models

Deep Learning models used in the paper "Developing a Victorious Strategy to the Second Strong Gravitational Lensing Data Challenge". We share the  Deep Learning models weights trained using II Strong Lensing Gravitational Challenge dataset in the configurations presented in table 1  from the aforementioned paper. We also release notebooks with examples of usage and adaptation to other datasets in <a href="https://github.com/cdebom/cast_lensfinder">github.com/cdebom/cast_lensfinder</a>. The file names are structure as follows: efn2_typeofentry_preprocessing_aug.tar.gz where typeofentry is the input format according to table 1 in the paper and preprocessing can be chpre or newpre  is the preprocessing used in the II SLGC, and the novel proposed preprocessing for the paper, respectively.

opencc-by-4.0Mar 2022View details →

zenodo36/100

Image dataset to train a deep learning model to decode Leetspeak obfuscated characters

The dataset contains an image database (18,981 images) that could be used to train a deep learning model to accurately detect characters. We have successfully used it to create a model that identifies characters encoded using LeetSpeak. The original dataset can be found in the Mondragon Unibertsitatea Repository -- https://gitlab.danz.eus/datasharing/ski4spam The training dataset consists of: - Alphabetic letters (a-z) written using different fonts and styles (regular, cursive, bold, cursive+bold) - Handwritten letters: English handwriting from the Chars74k dataset [2] which is available at http://www.ee.surrey.ac.uk/CVSSP/demos/chars74k/.

opencc-by-4.0Mar 2022View details →

zenodo36/100

Training dataset for "A deep learned nanowire segmentation model using synthetic data augmentation"

This image dataset contains synthetic structure images used for training the deep-learning based nanowire segmentation model presented in our work "A deep learned nanowire segmentation model using synthetic data augmentation" to be published in npj Computational materials. Detailed information can be found in the corresponding article.

opencc-by-4.0Apr 2022View details →

zenodo36/100

MoDL: Model-Based Deep Learning Dataset

This is the fully sampled raw dataset used in the MoDL paper acquired using the 3D T2-CUBE sequence at the University of Iowa. Please see the paper for the details. If you use this dataset then please consider citing this paper.  The python source code is also uploaded here for convenience.  Titled : MoDL: Model-Based Deep Learning Architecture for Inverse Problems by H.K. Aggarwal, M.P Mani, and Mathews Jacob in IEEE Transactions on Medical Imaging, 2019 IEEExplore Link: https://ieeexplore.ieee.org/document/8434321 Arxiv paper PDF Link: <a href="https://arxiv.org/abs/1712.02862">https://arxiv.org/abs/1712.02862</a> Python Source code link: <a href="https://github.com/hkaggarwal/modl">https://github.com/hkaggarwal/modl</a> Thank you

opencc-by-4.0Jan 2019View details →

zenodo36/100

Data for "Physically Based Deep Learning Framework to Model Intense Precipitation Events at Engineering Scales"

The dataset consists of high resolution (250 m) and low resolution (0.025 degree) climate model outputs in netCDF format. Each file contains data for one variable and one month. Low resolution files follow the naming scheme: montrealC_0025deg_200x200_ERA5_1m_YYYYMM_VAR.nc High resolution files follow the naming scheme: montrealC_250m_324x324_ERA5_TEB_100_noconv_YYYYMM_VAR.nc YYYYMM stands for the year (first 4 digits) and month (last 2 digits). _VAR indicates the variable contained in the file: <ul> <li>_UU700 stands for the east-west component of wind at a pressure level of 700 hPa (hourly frequency)</li> <li>_VV700 stands for the north-south component of wind at a pressure level of 700 hPa (hourly frequency)</li> <li>When _VAR is omitted, the variable is precipitation at 1-minute temporal resolution</li> </ul>

opencc-by-4.0Jun 2022View details →

zenodo36/100

Deep learning generates custom-made logistic regression models for explaining how breast cancer subtypes are classified

Breast cancer is the most frequently found cancer in women and the one most often subjected to genetic analysis. Nonetheless, it has been causing the largest number of women's cancer-related deaths. PAM50, the intrinsic subtype assay for breast cancer, is beneficial for diagnosis and stratified treatment but does not explain each subtype's mechanism. Nowadays, deep learning can predict the subtypes from genetic information more accurately than conventional statistical methods. However, the previous studies did not directly use deep learning to examine which genes associate with the subtypes. Ours is the first study on a deep-learning approach to reveal the mechanisms embedded in the PAM50-classified subtypes. We developed an explainable deep learning model called a point-wise linear model, which uses a meta-learning approach to generate a custom-made logistic regression model for each sample. Logistic regression is familiar to physicians and medical informatics researchers, and we can use it to analyze which genes are important for subtype prediction. The custom-made logistic regression models generated by the point-wise linear model for each subtype used the specific genes selected in other subtypes compared to the conventional logistic regression model: the overlap ratio is less than twenty percent. And analyzing the point-wise linear model's inner state, we found that the point-wise linear model used genes relevant to the cell cycle-related pathways. The results of this study suggest the potential of our explainable deep learning to play a vital role in cancer treatment.

opencc-by-4.0May 2021View details →

zenodo36/100

Data for: Deep Learning of Model- and Reanalysis- Based Precipitation and Pressure Mismatches over Europe

This study focuses on using UNet Convolutional Neural Networks to predict the spatiotemporal mismatches (errors) between TSMP-G2A model-based and COSMO-REA6 reanalysis-based precipitation and surface pressure over Europe. The following data are provided in this dataset: 1) The remapped and NetCDF-merged TSMP-G2A and COSMO-REA6 precipitation and surface pressure over the study area (EU-11 EUROCORDEX, ~0.11 degrees) for the years 1995-2017. Files: COSMO-REA6_PREPROCESSED.zip and TSMP_PREPROCESSED.zip 2) The actual and predicted spatiotemporal mismatch data for training, validation, and testing periods (1995-2017). Files: MISMATCH_ACTUAL.zip and MISMATCH_PREDICTED.zip   References for original TSMP-G2A and COSMO-REA6 data: TSMP-G2A: http://doi.org/10.17616/R31NJMGR COSMO-REA6: doi:10.1002/qj.2486, 2015

opencc-by-4.0Jul 2022View details →

zenodo36/100

Fig. 7 in An optimised YOLOv4 deep learning model for efficient malarial cell detection in thin blood smear images

Fig. 7 Confusion matrix of predictions by YOLOv4 models. YOLO,You Only Look Once (model)

opencc-by-4.0Apr 2024View details →

zenodo36/100

Simulations dataset and pre-trained models of "Deep learning in real-time on the astrophysical data obtained from the Čerenkov CTA Observatory" Ph.D. project

Ph.D. project datasets and models release, Deep learning in real-time on the astrophysical data obtained from the Čerenkov CTA Observatory.

opencc-by-4.0May 2024View details →

ScienceDex guides

Understand access before you commit

These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.

Compare curated datasets

Allen Brain Atlas

Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.

allen-brain-atlas

neuroscienceopenDocumentation, web resources, and API references are available online.

Last verified 2026-04-30Open record

Annotated Behaviour and Observability Dataset (ABODe)

ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.

abode-home-cage

behavioral-neuroscienceopenThe DataShare record exposes download links for annotations, documentation, license text, and the zipped per-snippet data directory.

Last verified 2026-04-30Open record

DANDI Archive for NWB datasets

DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.

dandi-nwb

electrophysiologyopenPublished Dandiset metadata and archive endpoints are available through the production DANDI API.

Last verified 2026-04-30Open record

International Brain Laboratory public data

The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.

ibl

behavioral-neuroscienceopenPublic sessions can be searched and loaded from the IBL public data server through ONE.

Last verified 2026-04-29Open record

OpenNeuro

OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.

openneuro

neuroscienceopenPublished datasets are available on demand over the internet.

Last verified 2026-04-29Open record