:memo: new. model metrics update
Browse files
README.md
CHANGED
@@ -58,16 +58,16 @@ We used heuristic parameters without separate hyperparameter tuning.
|
|
58 |
## Metrics
|
59 |
|
60 |
For the dev set we got those results
|
61 |
-
- size of the set: 700 examples from French Census /
|
62 |
-
- CER: 0.
|
63 |
-
- WER: 0.
|
64 |
-
- Loss: 0.
|
65 |
|
66 |
For the test set we got those results
|
67 |
- size of the set: 730 examples from French Census / 950 from our own dataset
|
68 |
-
- CER: 0.
|
69 |
-
- WER: 0.
|
70 |
-
- Loss: 0.
|
71 |
|
72 |
### How to use
|
73 |
|
|
|
58 |
## Metrics
|
59 |
|
60 |
For the dev set we got those results
|
61 |
+
- size of the set: 700 examples from French Census / 1600 from our own dataset
|
62 |
+
- CER: 0.0575
|
63 |
+
- WER: 0.1651
|
64 |
+
- Loss: 0.5768
|
65 |
|
66 |
For the test set we got those results
|
67 |
- size of the set: 730 examples from French Census / 950 from our own dataset
|
68 |
+
- CER: 0.09417
|
69 |
+
- WER: 0.23485
|
70 |
+
- Loss: 0.8700
|
71 |
|
72 |
### How to use
|
73 |
|