more readme

README.md

@@ -1,26 +1,32 @@
-# captioning
+### From Captions to Visual Concepts and Back ### 
+Code for detecting visual concepts in images.
+
 #### Installation Instructions ####
 0. Create directory, checkout code, caffe, coco-hooks
-
   ```shell
   git clone git@github.com:s-gupta/im2cap.git code
   git clone git@github.com:pdollar/coco.git coco
   ```
-
 0. Make caffe and pycaffe 
   ```shell
   git clone git@github.com:s-gupta/caffe.git caffe 
   cd caffe 
   git checkout mil
   make -j 16
   make pycaffe
+  cd
   ```
-### Get the data ###
-0. Get the COCO images
-
-0. Get the caffe image net models
-
-0. Get the pre-trained models 
+0. Get the COCO images, caffe imagenet models, pretrained models on COCO.
+  ``` shell
+  # Get the COCO images, splits, ground truth
+  wget http://www.cs.berkeley.edu/~sgupta/captions/data/data.tgz && tar -xf data.tgz
+  
+  # Get the caffe imagenet models 
+  wget http://www.cs.berkeley.edu/~sgupta/captions/data/caffe-data.tgz && tar -xf caffe-data.tgz
+  
+  # Get the pretrained models 
+  wget http://www.cs.berkeley.edu/~sgupta/captions/data/pretrained-coco.tgz && tar -xf pretrained-coco.tgz 
 
 
-### Testing the model ###
+#### Training, Testing the model ####
+```cd code``` and execute relevant commands from the file scripts/scripts_all.py 

output/README.md

@@ -0,0 +1,2 @@
+#### Directory for storing the prototxt files, snapshots and evaluations ####
+0. vgg - Contains the model that was trained on COCO for the CVPR 15 paper.