jruffle/nmf_samples_32d

NMF Model - samples mode - 32D

Pre-trained nmf model for transcriptomic data compression.

Details

• Mode: samples-centric compression
• Dimensions: 32
• Training data: TRACERx lung cancer transcriptomics
• Created: 2026-01-10T13:16:00.951114

Usage

import joblib
from huggingface_hub import snapshot_download

# Download model
local_dir = snapshot_download("jruffle/nmf_samples_32d")
model = joblib.load(f"{local_dir}/model.joblib")

# For classical models (PCA/UMAP):
# model contains: 'pca', 'umap', 'robust_scaler', 'gene_order'

# For TabPFN models (with UMAP reduction):
# model contains: 'tabpfn_embedding', 'umap_final', 'raw_embeddings', 'input_scaler', etc.
# Uses UMAP instead of PCA for non-linear dimension reduction of TabPFN embeddings