“WadoodAbdul”
commited on
Commit
·
28687f6
1
Parent(s):
88700dd
updated links
Browse files- src/about.py +4 -4
src/about.py
CHANGED
@@ -110,10 +110,10 @@ $$ Recall = COR / (COR + INC + MIS)$$
|
|
110 |
$$ f1score = 2 * (Prec * Rec) / (Prec + Rec)$$
|
111 |
|
112 |
Note:
|
113 |
-
1. Span-based approach here is equivalent to the 'Span Based Evaluation with Partial Overlap' in
|
114 |
-
2. Token-based approach here is equivalent to the 'Token Based Evaluation With Macro Average' in
|
115 |
|
116 |
-
Additional examples can be tested on the
|
117 |
|
118 |
## Datasets
|
119 |
The following datasets (test splits only) have been included in the evaluation.
|
@@ -229,7 +229,7 @@ Users are advised to approach the results with an understanding of the inherent
|
|
229 |
EVALUATION_QUEUE_TEXT = """
|
230 |
|
231 |
Currently, the benchmark supports evaluation for models hosted on the huggingface hub and of type encoder, decoder or gliner type models.
|
232 |
-
If your model needs a custom implementation, follow the steps outlined in the [
|
233 |
|
234 |
|
235 |
### Fields Explanation
|
|
|
110 |
$$ f1score = 2 * (Prec * Rec) / (Prec + Rec)$$
|
111 |
|
112 |
Note:
|
113 |
+
1. Span-based approach here is equivalent to the 'Span Based Evaluation with Partial Overlap' in [NER Metrics Showdown!](https://huggingface.co/spaces/wadood/ner_evaluation_metrics) and is equivalent to Partial Match ("Type") in the nervaluate python package.
|
114 |
+
2. Token-based approach here is equivalent to the 'Token Based Evaluation With Macro Average' in [NER Metrics Showdown!](https://huggingface.co/spaces/wadood/ner_evaluation_metrics)
|
115 |
|
116 |
+
Additional examples can be tested on the [NER Metrics Showdown!](https://huggingface.co/spaces/wadood/ner_evaluation_metrics) huggingface space.
|
117 |
|
118 |
## Datasets
|
119 |
The following datasets (test splits only) have been included in the evaluation.
|
|
|
229 |
EVALUATION_QUEUE_TEXT = """
|
230 |
|
231 |
Currently, the benchmark supports evaluation for models hosted on the huggingface hub and of type encoder, decoder or gliner type models.
|
232 |
+
If your model needs a custom implementation, follow the steps outlined in the [clinical_ner_benchmark](https://github.com/WadoodAbdul/clinical_ner_benchmark/blob/e66eb566f34e33c4b6c3e5258ac85aba42ec7894/docs/custom_model_implementation.md) repo or reach out to our team!
|
233 |
|
234 |
|
235 |
### Fields Explanation
|