ContainersNVIDIADeepSeek-R1NVIDIALlama-3.1-Nemotron-70B-InstructNVIDIAPyTorch
CollectionsNVIDIAOmniverse Kit (FB)NVIDIADeepStream SDKNVIDIAOmniverse Kit App Streaming
ModelsNVIDIAStyleGAN3 pretrained modelsNVIDIAPeopleNetNVIDIATrafficCamNet
ResourcesNVIDIARiva Skills Quick Start
Helm ChartsNVIDIAGPU OperatorNVIDIANVIDIA NIM Operator
NVIDIA
NVIDIA
Geneformer 10M
Model
NVIDIA
NVIDIA
Geneformer 10M

Geneformer is a tabular count model trained on sc-RNA from the Chan Zuckerberg Cell x Gene census

This model is backed by NVIDIA's Plus Plus (++) Promise
to learn more about the quality of the datasets used to train this model.
FieldResponse
Intended Application(s) & Domain(s):Generates representations of cells from genes
Model Type:Gene Expression Representation
Intended Users:This model is intended for academic and industrial computational biologists who study and perform experiments with single cell RNA sequencing data.
Outputs:Gene Expression Embeddings
Describe how the model works:Outputs a dense embedding based on masked genes and classification targets.
Technical Limitations:The model does not take into account the magnitude of gene expression, only the rank. The model is technically limited to only observing 1000-2000 genes rather than the entire set of possible targets (~20,000).
Verified to have met prescribed NVIDIA standards:Yes
Performance Metrics:* Reconstruction Error (Cross Entropy)
* Cell Type Classification
Potential Known Risks:The model may not work on cell types and genes unrelated to neuronal gene expression; model accuracy should be further researched for potential differences in population based on age, sex, and ethnicity.
Licensing:Apache 2.0