
Universal Cell Embeddings is a PyTorch implementation of the UCE model for embedding single-cell RNA sequencing datasets. It adds a cell embedding to each cell in an AnnData file.
Web appFree
Step 1: Install the dependencies with pip install -r requirements.txt.
Step 2: Prepare an h5ad file with counts and gene names as variable names.
Step 3: Run the eval_single_anndata.py script with your data path, output directory, species, model location and batch size.
Step 4: Use the nlayers option to run the larger model if you downloaded it.
Step 5: Read the embeddings from the X_uce slot of the output AnnData.
An h5ad file where the main matrix holds scRNA-seq counts and the variable names are gene names rather than Ensembl IDs.
In the obsm slot named X_uce of the output AnnData file, saved in your working directory.
No. Embeddings from the 33 layer model are not compatible with those from the 4 layer model.
It is per GPU. The README suggests 25 for the 33 layer model and 100 for the 4 layer model on an 80GB GPU.
MatGLGraph deep-learning library for materials science with pretrained interatomic potentials and property-prediction models
Arize AXArize AX is an AI-focused web app for developer tools workflows, giving users practical model-powered capabilities through a dedicated product or developer interface.