cmarkea
/

detr-layout-detection

Image Segmentation

Inference Endpoints

Model card Files Files and versions Community

Cyrile commited on Jul 30

Commit

c5278fa

•

1 Parent(s): bbb6212

Update README.md

Files changed (1) hide show

README.md +3 -2

README.md CHANGED Viewed

@@ -41,8 +41,9 @@ with torch.inference_mode():
     input_ids = img_proc(img, return_tensors='pt')
     output = model(**input_ids)
-segmentation_mask = img_proc.post_process_semantic_segmentation(
-    output,
     target_sizes=[img.size[::-1]]
 )

     input_ids = img_proc(img, return_tensors='pt')
     output = model(**input_ids)
+segmentation_mask = img_proc.post_process_segmentation(
+    out_seg,
+    threshold=0.5,
     target_sizes=[img.size[::-1]]
 )