Suggestion to improve Basic Usage readability.

Author: Narsil

README.md

@@ -30,7 +30,18 @@ Google and Stanford June 2020 paper [Sparse GPU Kernels for Deep Learning](https
 This would be even more general, as the sparsity pattern is not constrained, and the performance looks very good, with some smart ad hoc optimizations.
 
 ## Basic usage
-You can use the BlockSparseLinear drop in replacement for torch.nn.Linear in your own model.
+You can use the BlockSparseLinear drop in replacement for torch.nn.Linear in your own model:
+
+```python
+from pytorch_block_sparse import BlockSparseLinear
+
+...
+
+# self.fc = nn.Linear(1024, 256)
+self.fc = BlockSparseLinear(1024, 256, density=0.1)
+```
+
+### Secondary usage.
 
 Or you can use a utility called BlockSparseModelPatcher to modify easily an existing model before training it.(you cannot magically sparsify a trained existing model, you will need to train it from scratch)