Removing tensorflow as a dependency

README.md

@@ -262,7 +262,7 @@ For computational efficiency, we currently restricted the rationale to have maxi
 
 ## TensorBoard
 
-During training, TensorBoard logs are automatically saved to the same directory as the model checkpoints. To view TensorBoard logs, run `tensorboard --logdir=<dir>` where `<dir>` is the path to the checkpoint directory. Then navigate to [http://localhost:6006](http://localhost:6006).
+During training, TensorBoard logs are automatically saved to the same directory as the model checkpoints. To view TensorBoard logs, first install TensorFlow with `pip install tensorflow`. Then run `tensorboard --logdir=<dir>` where `<dir>` is the path to the checkpoint directory. Then navigate to [http://localhost:6006](http://localhost:6006).
 
 ## Results
 

environment.yml

@@ -20,7 +20,6 @@ dependencies:
   - scikit-learn>=0.22.2.post1
   - scipy>=1.4.1
   - tensorboardX>=2.0
-  - tensorflow>=1.13.1
   - torchvision>=0.5.0
   - tqdm>=4.45.0
   - pip:

setup.py

@@ -56,7 +56,6 @@
         'scipy==1.4.1',
         'sphinx>=3.1.2',
         'tensorboardX>=2.0',
-        'tensorflow>=1.13.1',
         'torch>=1.5.1',
         'tqdm>=4.45.0',
         'typed-argument-parser>=1.5.4'