YuDZEN commited on
Commit
da922da
1 Parent(s): 506a7e5

[update] README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -1
README.md CHANGED
@@ -18,4 +18,7 @@ Transcription data is `test.jsonl` and `train.jsonl` in the racine folder. The t
18
 
19
  ## Data Split
20
  The data is split into `train` and `test` sets.
21
- The `train` set contains 90% of the data and the `test` set contains 10% of the data.
 
 
 
 
18
 
19
  ## Data Split
20
  The data is split into `train` and `test` sets.
21
+ The `train` set contains 90% of the data and the `test` set contains 10% of the data.
22
+
23
+ ## data augmentation
24
+ We used data augmentation in order to increase the diversity of our training data without actually collecting new data (we have a very small dataset). By applying noise and volume, we can generate some training samples, which are named by "_NOISE" or "_VOLUME" after speakers' names.