Update README.md

README.md

@@ -215,4 +215,17 @@ for i, detail in enumerate(sentence_details, 1):
 
 print("="*80)
 
-```
+```
+
+If you want to quantize this model to save a lot of memory, you can use torchao.
+This is the config you would use if you wanted to run it on a laptop or small device
+
+```python
+from torchao.quantization import quantize_, Int8WeightOnlyConfig
+
+model.eval().to("cpu")
+
+# In-place: converts Linear layers to int8 weights
+quantize_(model, Int8WeightOnlyConfig())
+```
+