RaphaelOlivier commited on
Commit
c209c87
1 Parent(s): 7867261

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +7 -4
README.md CHANGED
@@ -1,9 +1,11 @@
 
1
  This dataset is a subset of [https://huggingface.co/datasets/librispeech_asr](LibriSpeech) that has been adversarially modified. It is designed to fool ASR models into predicting a target of our choosing instead of the correct output.
2
 
 
3
  The dataset contains several splits. Each split consists of the same utterances, modified with different types and amount of noise. 3 noises have been used:
4
- * Adversarial noise of radius $\epsilon=0.04$ (`adv_0.04` split)
5
- * Adversarial noise of radius $\epsilon=0.015$ (`adv_0.015` split)
6
- * Adversarial noise of radius $\epsilon=0.015$ combined with Room Impulse Response (RIR) noise (`adv_0.015_RIR` split)
7
 
8
  In addition we provide the original inputs (`natural` split)
9
 
@@ -17,6 +19,7 @@ For each noise we actually provide a split labeled with the original or "natural
17
  * `adv_0.015_RIR_nat_txt`
18
  * `adv_0.015_RIR_tgt_txt`
19
 
 
20
  You should evaluate your model on this dataset as you would evaluate it on LibriSpeech. Here is an example with Wav2Vec2
21
 
22
 
@@ -51,5 +54,5 @@ print("WER:", wer(result["text"], result["transcription"]))
51
 
52
  | "0.015 target text" | "0.015 natural text" | "0.04 target text" | "0.04 natural text"
53
  |---|---|---|---|
54
- | 58.2 | 108 | 58.2 | 49.5 |
55
 
 
1
+ # Description
2
  This dataset is a subset of [https://huggingface.co/datasets/librispeech_asr](LibriSpeech) that has been adversarially modified. It is designed to fool ASR models into predicting a target of our choosing instead of the correct output.
3
 
4
+ ## Splits
5
  The dataset contains several splits. Each split consists of the same utterances, modified with different types and amount of noise. 3 noises have been used:
6
+ * Adversarial noise of radius 0.04 (`adv_0.04` split)
7
+ * Adversarial noise of radius 0.015 (`adv_0.015` split)
8
+ * Adversarial noise of radius 0.015 combined with Room Impulse Response (RIR) noise (`adv_0.015_RIR` split)
9
 
10
  In addition we provide the original inputs (`natural` split)
11
 
 
19
  * `adv_0.015_RIR_nat_txt`
20
  * `adv_0.015_RIR_tgt_txt`
21
 
22
+ ## Usage
23
  You should evaluate your model on this dataset as you would evaluate it on LibriSpeech. Here is an example with Wav2Vec2
24
 
25
 
 
54
 
55
  | "0.015 target text" | "0.015 natural text" | "0.04 target text" | "0.04 natural text"
56
  |---|---|---|---|
57
+ | 58.2 | 108 | 49.5 | 108 |
58