update readme

README.md

@@ -15,6 +15,41 @@ widget:
 
 Trained for 4 epochs.
 
+Usage:
+
+```
+image_processor = AutoImageProcessor.from_pretrained("microsoft/dit-large")
+model = BeitForSemanticSegmentation.from_pretrained("jzju/dit-doclaynet")
+image = Image.open('img.png').convert('RGB')
+inputs = image_processor(images=image, return_tensors="pt")
+outputs = model(**inputs)
+# logits are of shape (batch_size, num_labels, height, width)
+logits = outputs.logits
+out = logits[0].detach()
+out.size()
+for i in range(11):
+    plt.imshow(out[i])
+    plt.show()
+```
+
+Labels:
+
+```
+1: Caption
+2: Footnote
+3: Formula
+4: List-item
+5: Page-footer
+6: Page-header
+7: Picture
+8: Section-header
+9: Table
+10: Text
+11: Title
+```
+
+Data label convert:
+
 ```
 model = BeitForSemanticSegmentation.from_pretrained("microsoft/dit-base", num_labels=11)
 ds = load_dataset("ds4sd/DocLayNet-v1.1")