Fix layout

README.md

@@ -94,7 +94,7 @@ python3 zipnn_decompress_path.py --path .
 ```
 
 Now just run the local version of the model.
-# Use a pipeline as a high-level helper
+### Use a pipeline as a high-level helper
 ```python
 from transformers import pipeline
 
@@ -105,7 +105,7 @@ pipe = pipeline("text-generation", model="PATH_TO_MODEL") # "." if in directory
 pipe(messages)
 ```
 
-# Load model directly
+### Load model directly
 ```python
 from transformers import AutoTokenizer, AutoModelForCausalLM