Add description to card metadata

#1
by julien-c HF staff - opened
Files changed (1) hide show
  1. README.md +38 -4
README.md CHANGED
@@ -1,6 +1,6 @@
1
  ---
2
  title: CER
3
- emoji: 🤗
4
  colorFrom: blue
5
  colorTo: red
6
  sdk: gradio
@@ -8,10 +8,44 @@ sdk_version: 3.0.2
8
  app_file: app.py
9
  pinned: false
10
  tags:
11
- - evaluate
12
- - metric
13
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
15
  # Metric Card for CER
16
 
17
  ## Metric description
 
1
  ---
2
  title: CER
3
+ emoji: 🤗
4
  colorFrom: blue
5
  colorTo: red
6
  sdk: gradio
 
8
  app_file: app.py
9
  pinned: false
10
  tags:
11
+ - evaluate
12
+ - metric
13
+ description: >-
14
+ Character error rate (CER) is a common metric of the performance of an
15
+ automatic speech recognition system.
16
+
17
+
18
+ CER is similar to Word Error Rate (WER), but operates on character instead of
19
+ word. Please refer to docs of WER for further information.
20
+
21
+
22
+ Character error rate can be computed as:
23
+
24
+
25
+ CER = (S + D + I) / N = (S + D + I) / (S + D + C)
26
+
27
+
28
+ where
29
 
30
+
31
+ S is the number of substitutions,
32
+
33
+ D is the number of deletions,
34
+
35
+ I is the number of insertions,
36
+
37
+ C is the number of correct characters,
38
+
39
+ N is the number of characters in the reference (N=S+D+C).
40
+
41
+
42
+ CER's output is not always a number between 0 and 1, in particular when there
43
+ is a high number of insertions. This value is often associated to the
44
+ percentage of characters that were incorrectly predicted. The lower the value,
45
+ the better the
46
+
47
+ performance of the ASR system with a CER of 0 being a perfect score.
48
+ ---
49
  # Metric Card for CER
50
 
51
  ## Metric description