lcampillos
commited on
Commit
•
a0c23f2
1
Parent(s):
fedc38c
Update README.md
Browse files
README.md
CHANGED
@@ -30,10 +30,10 @@ This named entity recognition model detects negation and speculation entities, a
|
|
30 |
- Speculated: speculated entity or event (e.g. *posiblemente **sobreviva***)
|
31 |
|
32 |
The model achieves the following results on the test set (when trained with the training and development set; results are averaged over 5 evaluation rounds):
|
33 |
-
- Precision: 0.
|
34 |
-
- Recall: 0.
|
35 |
-
- F1: 0.
|
36 |
-
- Accuracy: 0.
|
37 |
|
38 |
## Model description
|
39 |
|
@@ -100,24 +100,24 @@ The following hyperparameters were used during training:
|
|
100 |
- seed: we used different seeds for 5 evaluation rounds, and uploaded the model with the best results
|
101 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
102 |
- lr_scheduler_type: linear
|
103 |
-
- num_epochs:
|
104 |
|
105 |
|
106 |
### Training results (test set; average and standard deviation of 5 rounds with different seeds)
|
107 |
|
108 |
| Precision | Recall | F1 | Accuracy |
|
109 |
|:--------------:|:--------------:|:--------------:|:--------------:|
|
110 |
-
| 0.
|
111 |
|
112 |
|
113 |
**Results per class (test set; average and standard deviation of 5 rounds with different seeds)**
|
114 |
|
115 |
| Class | Precision | Recall | F1 | Support |
|
116 |
|:-----------:|:--------------:|:--------------:|:--------------:|:---------:|
|
117 |
-
| Neg_cue | 0.
|
118 |
-
| Negated | 0.
|
119 |
-
| Spec_cue | 0.
|
120 |
-
| Speculated | 0.
|
121 |
|
122 |
|
123 |
### Framework versions
|
|
|
30 |
- Speculated: speculated entity or event (e.g. *posiblemente **sobreviva***)
|
31 |
|
32 |
The model achieves the following results on the test set (when trained with the training and development set; results are averaged over 5 evaluation rounds):
|
33 |
+
- Precision: 0.838 (±0.003)
|
34 |
+
- Recall: 0.866 (±0.005)
|
35 |
+
- F1: 0.852 (±0.003)
|
36 |
+
- Accuracy: 0.986 (±0.001)
|
37 |
|
38 |
## Model description
|
39 |
|
|
|
100 |
- seed: we used different seeds for 5 evaluation rounds, and uploaded the model with the best results
|
101 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
102 |
- lr_scheduler_type: linear
|
103 |
+
- num_epochs: 8
|
104 |
|
105 |
|
106 |
### Training results (test set; average and standard deviation of 5 rounds with different seeds)
|
107 |
|
108 |
| Precision | Recall | F1 | Accuracy |
|
109 |
|:--------------:|:--------------:|:--------------:|:--------------:|
|
110 |
+
| 0.838 (±0.003) | 0.866 (±0.005) | 0.852 (±0.003) | 0.986 (±0.001) |
|
111 |
|
112 |
|
113 |
**Results per class (test set; average and standard deviation of 5 rounds with different seeds)**
|
114 |
|
115 |
| Class | Precision | Recall | F1 | Support |
|
116 |
|:-----------:|:--------------:|:--------------:|:--------------:|:---------:|
|
117 |
+
| Neg_cue | 0.945 (±0.004) | 0.961 (±0.002) | 0.953 (±0.003) | 2416 |
|
118 |
+
| Negated | 0.815 (±0.003) | 0.838 (±0.005) | 0.826 (±0.003) | 3064 |
|
119 |
+
| Spec_cue | 0.811 (±0.005) | 0.868 (±0.009) | 0.839 (±0.005) | 746 |
|
120 |
+
| Speculated | 0.685 (±0.009) | 0.719 (±0.016) | 0.701 (±0.008) | 993 |
|
121 |
|
122 |
|
123 |
### Framework versions
|