Update README.md

README.md

@@ -48,32 +48,6 @@ The following models are available from the InterScale Hugging Face repository.
 
 Here's a basic example: <!-- CONFIRM: function/argument names below follow scConcept's API as a template — adjust to InterScale's actual package interface. -->
 
-```python
-from interscale import InterScale
-import scanpy as sc
-import squidpy as sq
-
-# Load your spatial transcriptomics data
-adata = sc.read_h5ad("your_data.h5ad")
-
-# Normalize and log-transform (InterScale performs best with max normalized values of 4)
-sc.pp.normalize_total(adata)
-sc.pp.log1p(adata)
-
-# Build the spatial neighborhood graph
-# Use a radius-based graph for image-based data (e.g. Xenium, MERFISH, IMC),
-# or a hexagonal grid-based graph for sequencing-based data (e.g. Visium)
-sq.gr.spatial_neighbors(adata, coord_type="generic", radius=<INSERT RADIUS>)
-
-# Initialize InterScale and load a pretrained model
-model = InterScale(cache_dir="./cache/")
-model.load_config_and_model(model_name="<INSERT MODEL NAME>")
-
-# Extract local and global embeddings
-result = model.extract_embeddings(adata=adata)
-adata.obsm["X_interscale_local"] = result["local_emb"]
-adata.obsm["X_interscale_global"] = result["global_emb"]
-```
 
 ### Multi-scale downstream analysis