CellPLM

✅ Status: ready | Version: v1.0

Overview

Cell-centric (not gene-centric) architecture, highest batch throughput (batch_size=128), fast inference

!!! tip “When to choose CellPLM”

User needs fast inference, high throughput, million-cell scale processing, or cell-level (not gene-level) modeling

Specifications

Property

Value

Model

CellPLM

Version

v1.0

Tasks

embed, integrate

Modalities

RNA

Species

human

Gene IDs

symbol

Embedding Dim

512

GPU Required

Yes

Min VRAM

8 GB

Recommended VRAM

16 GB

CPU Fallback

Yes

Adapter Status

✅ ready

Quick Start

import omicverse as ov

# 1. Check model spec
info = ov.fm.describe_model("cellplm")

# 2. Profile your data
profile = ov.fm.profile_data("your_data.h5ad")

# 3. Validate compatibility
check = ov.fm.preprocess_validate("your_data.h5ad", "cellplm", "embed")

# 4. Run inference
result = ov.fm.run(
    task="embed",
    model_name="cellplm",
    adata_path="your_data.h5ad",
    output_path="output_cellplm.h5ad",
    device="auto",
)

# 5. Interpret results
metrics = ov.fm.interpret_results("output_cellplm.h5ad", task="embed")

Input Requirements

Requirement

Detail

Gene ID scheme

symbol

Preprocessing

Standard preprocessing. Model handles tokenization internally.

Data format

AnnData (.h5ad)

Batch key

.obs column for batch integration (optional)

Output Keys

After running ov.fm.run(), results are stored in the AnnData object:

Key

Location

Description

X_cellplm

adata.obsm

Cell embeddings (512-dim)

 
import scanpy as sc

adata = sc.read_h5ad("output_cellplm.h5ad")
embeddings = adata.obsm["X_cellplm"]  # shape: (n_cells, 512)

# Downstream analysis
sc.pp.neighbors(adata, use_rep="X_cellplm")
sc.tl.umap(adata)
sc.tl.leiden(adata, resolution=0.5)
sc.pl.umap(adata, color=["leiden"])

Resources

Hands-On Tutorial

For a step-by-step walkthrough with code, see the CellPLM Tutorial Notebook.