added model download

querying/README.md

@@ -54,8 +54,11 @@ success, r, t = q.solve(a < b, return_values=[i])
 
 
 ### Installation and Experiments
-`train_models.sh` downloads and trains several neural networks on common classification tasks.
+`train_models.sh` downloads and trains several neural networks on common classification tasks. This might take a long time.
 Many of the models use existing implementations and datasets. See section "Used Models" below for details.
+
+Alternatively you can also skip this step and [download](https://drive.google.com/file/d/17br5Y9Ta-dxoFqFw83Qtcpdhr1JLGfa_/view?usp=sharing) our provided models. We recommend placing the file `models.zip` in this folder and running `unzip models.zip` from the command line. Additionally you will need to run `bash download_data.sh` to download the GTSRB data. Note, that we can not provide you with Imagenet data samples. You will need to follow the instructions in `train_models.sh` to download them yourself.
+
 Once this has been run and the models are successfully trained  `run.py` can be used to perform the experiments from the evaluation section of the paper.
 `run_additional.py` performs the experiments found in the appendix of the paper.