Files changed (1) hide show
  1. README.md +44 -1
README.md CHANGED
@@ -3,4 +3,47 @@ license: mit
3
  base_model:
4
  - magistermilitum/tridis_HTR
5
  library_name: transformers
6
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  base_model:
4
  - magistermilitum/tridis_HTR
5
  library_name: transformers
6
+ language:
7
+ - la
8
+ ---
9
+ Base model: **magistermilitum/tridis_HTR v1**
10
+
11
+ Train Lines: ???
12
+
13
+ Eval Lines: ???
14
+
15
+ Test Lines: ???
16
+
17
+
18
+ Epochs: 14.1667 / 20
19
+
20
+ Eval CER: 0.0544
21
+
22
+ Test CER: 0.0622
23
+
24
+
25
+ Testresults with CERberus
26
+ | Metric | Value |
27
+ |----------------------------|---------|
28
+ | Character Error Rate | 6.22 |
29
+ | Number of Correct Characters| 186998 |
30
+ | Number of Substitutions | 5425 |
31
+ | Number of Insertions | 2933 |
32
+ | Number of Deletions | 3849 |
33
+ | Total Character Count | 196272 |
34
+ | Original Lines Count | 2288 |
35
+ | Discarded Lines Count | 0 |
36
+
37
+
38
+ Finetuned on an Anglicana-dataset, with mainly Middle Latin and few Middle English and Anglo-Norman text sources containing documents from:
39
+
40
+ - the Common Pleas (CP)
41
+ - the Justices (JUST)
42
+
43
+ from the English Legal Court Rolls.
44
+
45
+ The model has not been extensively tested.
46
+
47
+ Errors often occur in the Punctuation, which itself has an error rate of 44.44% which mostly consits of missed ‧ dots.
48
+
49
+ Potential biases are still to be identified.