---
title: "Geneformer V1 10M"
description: "Geneformer is a single-cell transcriptome foundation model designed to learn gene-network context from many human single-cell transcriptomes."
source: https://liatir.com/ai/models/ctheodoris-geneformer-v1-10m
---

# Geneformer V1 10M

Geneformer is a single-cell transcriptome foundation model designed to learn
gene-network context from many human single-cell transcriptomes.

## What it does

Geneformer can produce cell representations and supports downstream workflows
such as cell or gene classification, embedding extraction, and in-silico
perturbation analysis.

## Current status in Liatir

Geneformer V1 10M is installable as a signed Runtime Box. Liatir pins the
official V1 checkpoint and matching Genecorpus-30M dictionaries to one upstream
revision instead of following the repository's changing default model.

## Expected inputs

- Human AnnData `.h5ad` data with non-negative raw counts in `.X`.
- Ensembl gene IDs in `var["ensembl_id"]`, or Ensembl IDs as `var_names`.
- Optional `obs["n_counts"]`; Liatir computes totals from `.X` when absent.
- Optional `obs["filter_pass"]` to include only passing cells.

## Expected outputs

- Embedded AnnData with `obsm["X_geneformer"]`.
- Lightweight CSV preview and JSON summary.
- Input warnings, metrics, Jobs, Results, and reproducible provenance.

The current Liatir slice produces cell embeddings only. Gene-network inference,
classification, fine-tuning, and in-silico perturbation are not exposed by this
runtime.

## Hardware and installation

Measured CPU throughput was not viable for the current product target. The
published targets are macOS arm64 Metal and Linux x86_64 CUDA 12.9; Windows
x86_64 runs the Linux CUDA target through WSL2, which needs an NVIDIA GPU
visible inside WSL2. Liatir does not substitute a CPU or different CUDA target
when one of these exact environments is unavailable.

The input matrix is normalized per cell to 10,000 counts, scaled by the official
Genecorpus-30M gene medians, converted to the V1 rank-value encoding, and capped
at 2,048 gene tokens. Cell embeddings are mean-pooled from the second-to-last
hidden layer. The original input file is never modified.

## Official source

- [Geneformer on Hugging Face](https://huggingface.co/ctheodoris/Geneformer)
- [Geneformer documentation](https://geneformer.readthedocs.io/en/latest/getstarted.html)