Removed bug. Added thresholds.

Author: Tandon-A

README.md

@@ -70,7 +70,7 @@ Download the [Emotic dataset](https://drive.google.com/open?id=0B7sjGeF4f3FYQUVl
 
 You can also train and test models on Emotic dataset by using the [Colab_train_emotic notebook](https://github.com/Tandon-A/emotic/blob/master/Colab_train_emotic.ipynb). [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Tandon-A/emotic/blob/master/Colab_train_emotic.ipynb)
 
-The **trained models** to use for inference purposes are availble [here](https://drive.google.com/drive/folders/1e-JLA7V73CQD5pjTFCSWnKCmB0gCpV1D?usp=sharing). 
+The **trained models and thresholds** to use for inference purposes are availble [here](https://drive.google.com/drive/folders/1e-JLA7V73CQD5pjTFCSWnKCmB0gCpV1D?usp=sharing). 
 
 ## Results 
 

inference.py

@@ -5,6 +5,8 @@
 import torch 
 from torchvision import transforms
 
+from emotic import Emotic 
+
 
 def process_images(context_norm, body_norm, image_context_path=None, image_context=None, image_body=None, bbox=None):
   ''' Prepare context and body image. 

test.py

@@ -9,6 +9,7 @@
 import torchvision.models as models
 from torchvision import transforms
 
+from emotic import Emotic 
 from emotic_dataset import Emotic_PreDataset