clarifications

README.md

@@ -16,7 +16,7 @@ A family of simple yet powerful deep neural networks for visual neuroscience. Wh
 
 ## Available Models
 
-*(click on model names to download the weights of ImageNet-trained models)*
+*(Click on model names to download the weights of ImageNet-trained models. Note that you do not need to untar them to use during testing.)*
 
 | Name     | Description                                                              |
 | -------- | ------------------------------------------------------------------------ |
@@ -39,7 +39,7 @@ Add `-o <path to save features>` if you want model responses to be saved somepla
 
 1. [Get ImageNet](https://github.com/facebook/fb.resnet.torch/blob/master/INSTALL.md#download-the-imagenet-dataset) if you don't have it yet. If you do, note that validation images need to be put in separate folders, just like train data. Follow the instructions in that link to do so easily.
 
-2. `python run.py train - --model Z`. On a single Titan X, it will train for about 20 hours. **NOTE** the extra `-` between `train` and `--model`!
+2. `python run.py train - --model Z --workers 20`. On a single Titan X, it will train for about 20 hours. Models R and S will require 2 GPUs. **NOTE** the extra `-` between `train` and `--model`!
 
 
 # Details