add int8

README.md

@@ -75,6 +75,13 @@ It is possible to export your model to the [OpenVINO](https://docs.openvino.ai/2
 optimum-cli export openvino --model distilbert-base-uncased-finetuned-sst-2-english ov_distilbert
 ```
 
+To apply int8 quantization on your model weights and keep the activations in floating point precision, you can add `--int8`:
+
+```plain
+optimum-cli export openvino --model distilbert-base-uncased-finetuned-sst-2-english --int8 ov_distilbert
+```
+
+
 #### Inference:
 
 To load a model and run inference with OpenVINO Runtime, you can just replace your `AutoModelForXxx` class with the corresponding `OVModelForXxx` class.