Update README.md

README.md

@@ -19,6 +19,10 @@ This model was trained on the datasets pc-nih-rsna-siim-vin at a 512x512 resolut
 
 Here is how to use this model to classify an image of xray:
 
+Note: Each pretrained model has 18 outputs. The `all` model has every output trained. However, for the other weights some targets are not trained and will predict randomly becuase they do not exist in the training dataset. The only valid outputs are listed in the field `{dataset}.pathologies` on the dataset that corresponds to the weights. 
+
+
+Benchmarks of the modes are here: [BENCHMARKS.md](https://github.com/mlmed/torchxrayvision/blob/master/BENCHMARKS.md)
 ```python
 import urllib.request