diff --git a/README.md b/README.md
index b08173b62cdfa1d89d45edbce65a84dd9c6dc750..91e9815bd17e873663bb2ac92129e23a272d010e 100644
--- a/README.md
+++ b/README.md
@@ -96,7 +96,7 @@ Here is a short overview of the stages:
 - ```stage=2```: 
   - counts frequencies of used codebook entries per speaker
   - if VERBOSE is true this stage also generates log-files
-  - **if you need to extract features with a CPU, set ```device = torch.device('cpu')``` in the script ```local/codebook_freqs.py``` (default is ```device = torch.device('cuda')```)**
+  - **if you need to extract features with a CPU, in the script ```local/codebook_freqs.py``` set ```device = torch.device('cpu')``` (default is ```device = torch.device('cuda')```)**
   - resulting files are stored in ```exp_*/logs/```, ```exp_*/numpy/``` and```exp_*/txt/```
 - ```stage=3```: 
   - prepares and stores a similarty matrix in the folder ```exp_*/numpy/```