Update README.md
Browse files
README.md
CHANGED
|
@@ -1,93 +1,1187 @@
|
|
| 1 |
---
|
| 2 |
-
library_name: transformers
|
| 3 |
license: mit
|
| 4 |
base_model: almanach/camembertv2-base
|
| 5 |
-
tags:
|
| 6 |
-
- generated_from_trainer
|
| 7 |
metrics:
|
| 8 |
- precision
|
| 9 |
- recall
|
| 10 |
- f1
|
| 11 |
- accuracy
|
| 12 |
model-index:
|
| 13 |
-
- name:
|
| 14 |
results: []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
---
|
| 16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
```
|
| 18 |
-
{'LOC': {'precision': 0.9510338498083464,
|
| 19 |
-
'recall': 0.9654366094263792,
|
| 20 |
-
'f1': 0.9581811094289677,
|
| 21 |
-
'number': 54740},
|
| 22 |
-
'MISC': {'precision': 0.8600569108290437,
|
| 23 |
-
'recall': 0.7587510224804671,
|
| 24 |
-
'f1': 0.806234077626255,
|
| 25 |
-
'number': 35453},
|
| 26 |
-
'O': {'precision': 0.9909218534126304,
|
| 27 |
-
'recall': 0.9936490359966582,
|
| 28 |
-
'f1': 0.9922835708676133,
|
| 29 |
-
'number': 805547},
|
| 30 |
-
'ORG': {'precision': 0.8822008564272441,
|
| 31 |
-
'recall': 0.921045972163644,
|
| 32 |
-
'f1': 0.901205018157808,
|
| 33 |
-
'number': 11855},
|
| 34 |
-
'PER': {'precision': 0.973038794785731,
|
| 35 |
-
'recall': 0.9823632323041278,
|
| 36 |
-
'f1': 0.9776787815093096,
|
| 37 |
-
'number': 63447},
|
| 38 |
-
'overall_precision': 0.9818586631680195,
|
| 39 |
-
'overall_recall': 0.9818586631680195,
|
| 40 |
-
'overall_f1': 0.9818586631680195,
|
| 41 |
-
'overall_accuracy': 0.9818586631680195}
|
| 42 |
-
```
|
| 43 |
|
| 44 |
-
# camembertv2-base-frenchNER_3entities
|
| 45 |
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
- Precision: 0.9822
|
| 50 |
-
- Recall: 0.9822
|
| 51 |
-
- F1: 0.9822
|
| 52 |
-
- Accuracy: 0.9822
|
| 53 |
|
| 54 |
-
##
|
| 55 |
|
| 56 |
-
|
| 57 |
|
| 58 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
-
More information needed
|
| 61 |
|
| 62 |
-
## Training and evaluation data
|
| 63 |
|
| 64 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 65 |
|
| 66 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
|
| 68 |
-
###
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 69 |
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 76 |
-
- lr_scheduler_type: linear
|
| 77 |
-
- num_epochs: 3
|
| 78 |
|
| 79 |
-
###
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 86 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
|
| 88 |
-
###
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 89 |
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
- Datasets 2.21.0
|
| 93 |
-
- Tokenizers 0.20.1
|
|
|
|
| 1 |
---
|
|
|
|
| 2 |
license: mit
|
| 3 |
base_model: almanach/camembertv2-base
|
|
|
|
|
|
|
| 4 |
metrics:
|
| 5 |
- precision
|
| 6 |
- recall
|
| 7 |
- f1
|
| 8 |
- accuracy
|
| 9 |
model-index:
|
| 10 |
+
- name: NERmembert2-4entities
|
| 11 |
results: []
|
| 12 |
+
datasets:
|
| 13 |
+
- CATIE-AQ/frenchNER_4entities
|
| 14 |
+
language:
|
| 15 |
+
- fr
|
| 16 |
+
widget:
|
| 17 |
+
- text: "Le dévoilement du logo officiel des JO s'est déroulé le 21 octobre 2019 au Grand Rex. Ce nouvel emblème et cette nouvelle typographie ont été conçus par le designer Sylvain Boyer avec les agences Royalties & Ecobranding. Rond, il rassemble trois symboles : une médaille d'or, la flamme olympique et Marianne, symbolisée par un visage de femme mais privée de son bonnet phrygien caractéristique. La typographie dessinée fait référence à l'Art déco, mouvement artistique des années 1920, décennie pendant laquelle ont eu lieu pour la dernière fois les Jeux olympiques à Paris en 1924. Pour la première fois, ce logo sera unique pour les Jeux olympiques et les Jeux paralympiques."
|
| 18 |
+
library_name: transformers
|
| 19 |
+
pipeline_tag: token-classification
|
| 20 |
+
co2_eq_emissions: 25.5
|
| 21 |
---
|
| 22 |
|
| 23 |
+
|
| 24 |
+
# NERmembert-large-4entities
|
| 25 |
+
|
| 26 |
+
## Model Description
|
| 27 |
+
|
| 28 |
+
We present **NERmembert2-4entities**, which is a [CamemBERT v2 base](https://huggingface.co/almanach/camembertv2-base) fine-tuned for the Name Entity Recognition task for the French language on four French NER datasets for 4 entities (LOC, PER, ORG, MISC).
|
| 29 |
+
All these datasets were concatenated and cleaned into a single dataset that we called [frenchNER_4entities](https://huggingface.co/datasets/CATIE-AQ/frenchNER_4entities).
|
| 30 |
+
There are a total of **384,773** rows, of which **328,757** are for training, **24,131** for validation and **31,885** for testing.
|
| 31 |
+
Our methodology is described in a blog post available in [English](https://blog.vaniila.ai/en/NER_en/) or [French](https://blog.vaniila.ai/NER/).
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
## Evaluation results
|
| 35 |
+
|
| 36 |
+
The evaluation was carried out using the [**evaluate**](https://pypi.org/project/evaluate/) python package.
|
| 37 |
+
|
| 38 |
+
### frenchNER_4entities
|
| 39 |
+
|
| 40 |
+
For space reasons, we show only the F1 of the different models. You can see the full results below the table.
|
| 41 |
+
|
| 42 |
+
<table>
|
| 43 |
+
<thead>
|
| 44 |
+
<tr>
|
| 45 |
+
<th><br>Model</th>
|
| 46 |
+
<th><br>PER</th>
|
| 47 |
+
<th><br>LOC</th>
|
| 48 |
+
<th><br>ORG</th>
|
| 49 |
+
<th><br>MISC</th>
|
| 50 |
+
</tr>
|
| 51 |
+
</thead>
|
| 52 |
+
<tbody>
|
| 53 |
+
<tr>
|
| 54 |
+
<td rowspan="1"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 55 |
+
<td><br>0.971</td>
|
| 56 |
+
<td><br>0.947</td>
|
| 57 |
+
<td><br>0.902</td>
|
| 58 |
+
<td><br>0.663</td>
|
| 59 |
+
</tr>
|
| 60 |
+
<tr>
|
| 61 |
+
<td rowspan="1"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 62 |
+
<td><br>0.974</td>
|
| 63 |
+
<td><br>0.948</td>
|
| 64 |
+
<td><br>0.892</td>
|
| 65 |
+
<td><br>0.658</td>
|
| 66 |
+
</tr>
|
| 67 |
+
<tr>
|
| 68 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 69 |
+
<td><br>0.978</td>
|
| 70 |
+
<td><br>0.958</td>
|
| 71 |
+
<td><br>0.903</td>
|
| 72 |
+
<td><br>0.814</td>
|
| 73 |
+
</tr>
|
| 74 |
+
<tr>
|
| 75 |
+
<td rowspan="1"><br>NERmembert2-4entities (this model) (111M)</td>
|
| 76 |
+
<td><br>0.978</td>
|
| 77 |
+
<td><br>0.958</td>
|
| 78 |
+
<td><br>0.901</td>
|
| 79 |
+
<td><br>0.806</td>
|
| 80 |
+
</tr>
|
| 81 |
+
<tr>
|
| 82 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 83 |
+
<td><br>0.979</td>
|
| 84 |
+
<td><br>0.961</td>
|
| 85 |
+
<td><br>0.915</td>
|
| 86 |
+
<td><br>0.812</td>
|
| 87 |
+
</tr>
|
| 88 |
+
<tr>
|
| 89 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 90 |
+
<td><br><b>0.982</b></td>
|
| 91 |
+
<td><br><b>0.964</b></td>
|
| 92 |
+
<td><br><b>0.919</b></td>
|
| 93 |
+
<td><br><b>0.834</b></td>
|
| 94 |
+
</tr>
|
| 95 |
+
</tbody>
|
| 96 |
+
</table>
|
| 97 |
+
|
| 98 |
+
|
| 99 |
+
<details>
|
| 100 |
+
<summary>Full results</summary>
|
| 101 |
+
<table>
|
| 102 |
+
<thead>
|
| 103 |
+
<tr>
|
| 104 |
+
<th><br>Model</th>
|
| 105 |
+
<th><br>Metrics</th>
|
| 106 |
+
<th><br>PER</th>
|
| 107 |
+
<th><br>LOC</th>
|
| 108 |
+
<th><br>ORG</th>
|
| 109 |
+
<th><br>MISC</th>
|
| 110 |
+
<th><br>O</th>
|
| 111 |
+
<th><br>Overall</th>
|
| 112 |
+
</tr>
|
| 113 |
+
</thead>
|
| 114 |
+
<tbody>
|
| 115 |
+
<tr>
|
| 116 |
+
<td rowspan="3"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 117 |
+
<td><br>Precision</td>
|
| 118 |
+
<td><br>0.952</td>
|
| 119 |
+
<td><br>0.924</td>
|
| 120 |
+
<td><br>0.870</td>
|
| 121 |
+
<td><br>0.845</td>
|
| 122 |
+
<td><br>0.986</td>
|
| 123 |
+
<td><br>0.976</td>
|
| 124 |
+
</tr>
|
| 125 |
+
<tr>
|
| 126 |
+
<td><br>Recall</td>
|
| 127 |
+
<td><br>0.990</td>
|
| 128 |
+
<td><br>0.972</td>
|
| 129 |
+
<td><br>0.938</td>
|
| 130 |
+
<td><br>0.546</td>
|
| 131 |
+
<td><br>0.992</td>
|
| 132 |
+
<td><br>0.976</td>
|
| 133 |
+
</tr>
|
| 134 |
+
<tr>
|
| 135 |
+
<td>F1</td>
|
| 136 |
+
<td><br>0.971</td>
|
| 137 |
+
<td><br>0.947</td>
|
| 138 |
+
<td><br>0.902</td>
|
| 139 |
+
<td><br>0.663</td>
|
| 140 |
+
<td><br>0.989</td>
|
| 141 |
+
<td><br>0.976</td>
|
| 142 |
+
</tr>
|
| 143 |
+
<tr>
|
| 144 |
+
<td rowspan="3"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 145 |
+
<td><br>Precision</td>
|
| 146 |
+
<td><br>0.962</td>
|
| 147 |
+
<td><br>0.933</td>
|
| 148 |
+
<td><br>0.857</td>
|
| 149 |
+
<td><br>0.830</td>
|
| 150 |
+
<td><br>0.985</td>
|
| 151 |
+
<td><br>0.976</td>
|
| 152 |
+
</tr>
|
| 153 |
+
<tr>
|
| 154 |
+
<td><br>Recall</td>
|
| 155 |
+
<td><br>0.987</td>
|
| 156 |
+
<td><br>0.963</td>
|
| 157 |
+
<td><br>0.930</td>
|
| 158 |
+
<td><br>0.545</td>
|
| 159 |
+
<td><br>0.993</td>
|
| 160 |
+
<td><br>0.976</td>
|
| 161 |
+
</tr>
|
| 162 |
+
<tr>
|
| 163 |
+
<td>F1</td>
|
| 164 |
+
<td><br>0.974</td>
|
| 165 |
+
<td><br>0.948</td>
|
| 166 |
+
<td><br>0.892</td>
|
| 167 |
+
<td><br>0.658</td>
|
| 168 |
+
<td><br>0.989</td>
|
| 169 |
+
<td><br>0.976</td>
|
| 170 |
+
</tr>
|
| 171 |
+
<tr>
|
| 172 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 173 |
+
<td><br>Precision</td>
|
| 174 |
+
<td><br>0.973</td>
|
| 175 |
+
<td><br>0.951</td>
|
| 176 |
+
<td><br>0.888</td>
|
| 177 |
+
<td><br>0.850</td>
|
| 178 |
+
<td><br>0.993</td>
|
| 179 |
+
<td><br>0.984</td>
|
| 180 |
+
</tr>
|
| 181 |
+
<tr>
|
| 182 |
+
<td><br>Recall</td>
|
| 183 |
+
<td><br>0.983</td>
|
| 184 |
+
<td><br>0.964</td>
|
| 185 |
+
<td><br>0.918</td>
|
| 186 |
+
<td><br>0.781</td>
|
| 187 |
+
<td><br>0.993</td>
|
| 188 |
+
<td><br>0.984</td>
|
| 189 |
+
</tr>
|
| 190 |
+
<tr>
|
| 191 |
+
<td>F1</td>
|
| 192 |
+
<td><br>0.978</td>
|
| 193 |
+
<td><br>0.958</td>
|
| 194 |
+
<td><br>0.903</td>
|
| 195 |
+
<td><br>0.814</td>
|
| 196 |
+
<td><br>0.993</td>
|
| 197 |
+
<td><br>0.984</td>
|
| 198 |
+
</tr>
|
| 199 |
+
<tr>
|
| 200 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert2-4entities">NERmembert2-4entities (this model) (111M)</a></td>
|
| 201 |
+
<td><br>Precision</td>
|
| 202 |
+
<td><br>TODO</td>
|
| 203 |
+
<td><br>TODO</td>
|
| 204 |
+
<td><br>TODO</td>
|
| 205 |
+
<td><br>TODO</td>
|
| 206 |
+
<td><br>TODO</td>
|
| 207 |
+
<td><br>TODO</td>
|
| 208 |
+
</tr>
|
| 209 |
+
<tr>
|
| 210 |
+
<td><br>Recall</td>
|
| 211 |
+
<td><br>TODO</td>
|
| 212 |
+
<td><br>TODO</td>
|
| 213 |
+
<td><br>TODO</td>
|
| 214 |
+
<td><br>TODO</td>
|
| 215 |
+
<td><br>TODO</td>
|
| 216 |
+
<td><br>TODO</td>
|
| 217 |
+
</tr>
|
| 218 |
+
<tr>
|
| 219 |
+
<td>F1</td>
|
| 220 |
+
<td><br>TODO</td>
|
| 221 |
+
<td><br>TODO</td>
|
| 222 |
+
<td><br>TODO</td>
|
| 223 |
+
<td><br>TODO</td>
|
| 224 |
+
<td><br>TODO</td>
|
| 225 |
+
<td><br>TODO</td>
|
| 226 |
+
</tr>
|
| 227 |
+
<tr>
|
| 228 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 229 |
+
<td><br>Precision</td>
|
| 230 |
+
<td><br>TODO</td>
|
| 231 |
+
<td><br>TODO</td>
|
| 232 |
+
<td><br>TODO</td>
|
| 233 |
+
<td><br>TODO</td>
|
| 234 |
+
<td><br>TODO</td>
|
| 235 |
+
<td><br>TODO</td>
|
| 236 |
+
</tr>
|
| 237 |
+
<tr>
|
| 238 |
+
<td><br>Recall</td>
|
| 239 |
+
<td><br>TODO</td>
|
| 240 |
+
<td><br>TODO</td>
|
| 241 |
+
<td><br>TODO</td>
|
| 242 |
+
<td><br>TODO</td>
|
| 243 |
+
<td><br>TODO</td>
|
| 244 |
+
<td><br>TODO</td>
|
| 245 |
+
</tr>
|
| 246 |
+
<tr>
|
| 247 |
+
<td>F1</td>
|
| 248 |
+
<td><br>TODO</td>
|
| 249 |
+
<td><br>TODO</td>
|
| 250 |
+
<td><br>TODO</td>
|
| 251 |
+
<td><br>TODO</td>
|
| 252 |
+
<td><br>TODO</td>
|
| 253 |
+
<td><br>TODO</td>
|
| 254 |
+
</tr>
|
| 255 |
+
<tr>
|
| 256 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 257 |
+
<td><br>Precision</td>
|
| 258 |
+
<td><br>0.977</td>
|
| 259 |
+
<td><br>0.961</td>
|
| 260 |
+
<td><br>0.896</td>
|
| 261 |
+
<td><br>0.872</td>
|
| 262 |
+
<td><br>0.993</td>
|
| 263 |
+
<td><br>0.986</td>
|
| 264 |
+
</tr>
|
| 265 |
+
<tr>
|
| 266 |
+
<td><br>Recall</td>
|
| 267 |
+
<td><br>0.987</td>
|
| 268 |
+
<td><br>0.966</td>
|
| 269 |
+
<td><br>0.943</td>
|
| 270 |
+
<td><br>0.798</td>
|
| 271 |
+
<td><br>0.995</td>
|
| 272 |
+
<td><br>0.986</td>
|
| 273 |
+
</tr>
|
| 274 |
+
<tr>
|
| 275 |
+
<td>F1</td>
|
| 276 |
+
<td><br><b>0.982</b></td>
|
| 277 |
+
<td><br><b>0.964</b></td>
|
| 278 |
+
<td><br><b>0.919</b></td>
|
| 279 |
+
<td><br><b>0.834</b></td>
|
| 280 |
+
<td><br><b>0.994</b></td>
|
| 281 |
+
<td><br><b>0.986</b></td>
|
| 282 |
+
</tr>
|
| 283 |
+
</tbody>
|
| 284 |
+
</table>
|
| 285 |
+
</details>
|
| 286 |
+
|
| 287 |
+
In detail:
|
| 288 |
+
|
| 289 |
+
### multiconer
|
| 290 |
+
|
| 291 |
+
For space reasons, we show only the F1 of the different models. You can see the full results below the table.
|
| 292 |
+
|
| 293 |
+
<table>
|
| 294 |
+
<thead>
|
| 295 |
+
<tr>
|
| 296 |
+
<th><br>Model</th>
|
| 297 |
+
<th><br>PER</th>
|
| 298 |
+
<th><br>LOC</th>
|
| 299 |
+
<th><br>ORG</th>
|
| 300 |
+
<th><br>MISC</th>
|
| 301 |
+
</tr>
|
| 302 |
+
</thead>
|
| 303 |
+
<tbody>
|
| 304 |
+
<tr>
|
| 305 |
+
<td rowspan="1"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 306 |
+
<td><br>0.940</td>
|
| 307 |
+
<td><br>0.761</td>
|
| 308 |
+
<td><br>0.723</td>
|
| 309 |
+
<td><br>0.560</td>
|
| 310 |
+
</tr>
|
| 311 |
+
<tr>
|
| 312 |
+
<td rowspan="1"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 313 |
+
<td><br>0.921</td>
|
| 314 |
+
<td><br>0.748</td>
|
| 315 |
+
<td><br>0.694</td>
|
| 316 |
+
<td><br>0.530</td>
|
| 317 |
+
</tr>
|
| 318 |
+
<tr>
|
| 319 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 320 |
+
<td><br>0.960</td>
|
| 321 |
+
<td><br>0.890</td>
|
| 322 |
+
<td><br>0.867</td>
|
| 323 |
+
<td><br>0.852</td>
|
| 324 |
+
</tr>
|
| 325 |
+
<tr>
|
| 326 |
+
<td rowspan="1"><br>NERmembert2-4entities (this model) (111M)</td>
|
| 327 |
+
<td><br>TODO</td>
|
| 328 |
+
<td><br>TODO</td>
|
| 329 |
+
<td><br>TODO</td>
|
| 330 |
+
<td><br>TODO</td>
|
| 331 |
+
</tr>
|
| 332 |
+
<tr>
|
| 333 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 334 |
+
<td><br>TODO</td>
|
| 335 |
+
<td><br>TODO</td>
|
| 336 |
+
<td><br>TODO</td>
|
| 337 |
+
<td><br>TODO</td>
|
| 338 |
+
</tr>
|
| 339 |
+
<tr>
|
| 340 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 341 |
+
<td><br><b>0.969</b></td>
|
| 342 |
+
<td><br><b>0.919</b></td>
|
| 343 |
+
<td><br><b>0.904</b></td>
|
| 344 |
+
<td><br><b>0.864</b></td>
|
| 345 |
+
</tr>
|
| 346 |
+
</tbody>
|
| 347 |
+
</table>
|
| 348 |
+
|
| 349 |
+
<details>
|
| 350 |
+
<summary>Full results</summary>
|
| 351 |
+
<table>
|
| 352 |
+
<thead>
|
| 353 |
+
<tr>
|
| 354 |
+
<th><br>Model</th>
|
| 355 |
+
<th><br>Metrics</th>
|
| 356 |
+
<th><br>PER</th>
|
| 357 |
+
<th><br>LOC</th>
|
| 358 |
+
<th><br>ORG</th>
|
| 359 |
+
<th><br>MISC</th>
|
| 360 |
+
<th><br>O</th>
|
| 361 |
+
<th><br>Overall</th>
|
| 362 |
+
</tr>
|
| 363 |
+
</thead>
|
| 364 |
+
<tbody>
|
| 365 |
+
<tr>
|
| 366 |
+
<td rowspan="3"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 367 |
+
<td><br>Precision</td>
|
| 368 |
+
<td><br>0.908</td>
|
| 369 |
+
<td><br>0.717</td>
|
| 370 |
+
<td><br>0.753</td>
|
| 371 |
+
<td><br>0.620</td>
|
| 372 |
+
<td><br>0.936</td>
|
| 373 |
+
<td><br>0.889</td>
|
| 374 |
+
</tr>
|
| 375 |
+
<tr>
|
| 376 |
+
<td><br>Recall</td>
|
| 377 |
+
<td><br>0.975</td>
|
| 378 |
+
<td><br>0.811</td>
|
| 379 |
+
<td><br>0.696</td>
|
| 380 |
+
<td><br>0.511</td>
|
| 381 |
+
<td><br>0.938</td>
|
| 382 |
+
<td><br>0.889</td>
|
| 383 |
+
</tr>
|
| 384 |
+
<tr>
|
| 385 |
+
<td>F1</td>
|
| 386 |
+
<td><br>0.940</td>
|
| 387 |
+
<td><br>0.761</td>
|
| 388 |
+
<td><br>0.723</td>
|
| 389 |
+
<td><br>0.560</td>
|
| 390 |
+
<td><br>0.937</td>
|
| 391 |
+
<td><br>0.889</td>
|
| 392 |
+
</tr>
|
| 393 |
+
<tr>
|
| 394 |
+
<td rowspan="3"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 395 |
+
<td><br>Precision</td>
|
| 396 |
+
<td><br>0.885</td>
|
| 397 |
+
<td><br>0.738</td>
|
| 398 |
+
<td><br>0.737</td>
|
| 399 |
+
<td><br>0.589</td>
|
| 400 |
+
<td><br>0.928</td>
|
| 401 |
+
<td><br>0.881</td>
|
| 402 |
+
</tr>
|
| 403 |
+
<tr>
|
| 404 |
+
<td><br>Recall</td>
|
| 405 |
+
<td><br>0.960</td>
|
| 406 |
+
<td><br>0.759</td>
|
| 407 |
+
<td><br>0.655</td>
|
| 408 |
+
<td><br>0.482</td>
|
| 409 |
+
<td><br>0.939</td>
|
| 410 |
+
<td><br>0.881</td>
|
| 411 |
+
</tr>
|
| 412 |
+
<tr>
|
| 413 |
+
<td>F1</td>
|
| 414 |
+
<td><br>0.921</td>
|
| 415 |
+
<td><br>0.748</td>
|
| 416 |
+
<td><br>0.694</td>
|
| 417 |
+
<td><br>0.530</td>
|
| 418 |
+
<td><br>0.934</td>
|
| 419 |
+
<td><br>0.881</td>
|
| 420 |
+
</tr>
|
| 421 |
+
<tr>
|
| 422 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 423 |
+
<td><br>Precision</td>
|
| 424 |
+
<td><br>0.954</td>
|
| 425 |
+
<td><br>0.893</td>
|
| 426 |
+
<td><br>0.851</td>
|
| 427 |
+
<td><br>0.849</td>
|
| 428 |
+
<td><br>0.979</td>
|
| 429 |
+
<td><br>0.954</td>
|
| 430 |
+
</tr>
|
| 431 |
+
<tr>
|
| 432 |
+
<td><br>Recall</td>
|
| 433 |
+
<td><br>0.967</td>
|
| 434 |
+
<td><br>0.887</td>
|
| 435 |
+
<td><br>0.883</td>
|
| 436 |
+
<td><br>0.855</td>
|
| 437 |
+
<td><br>0.974</td>
|
| 438 |
+
<td><br>0.954</td>
|
| 439 |
+
</tr>
|
| 440 |
+
<tr>
|
| 441 |
+
<td>F1</td>
|
| 442 |
+
<td><br>0.960</td>
|
| 443 |
+
<td><br>0.890</td>
|
| 444 |
+
<td><br>0.867</td>
|
| 445 |
+
<td><br>0.852</td>
|
| 446 |
+
<td><br>0.977</td>
|
| 447 |
+
<td><br>0.954</td>
|
| 448 |
+
</tr>
|
| 449 |
+
<tr>
|
| 450 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert2-4entities">NERmembert2-4entities (this model) (111M)</a></td>
|
| 451 |
+
<td><br>Precision</td>
|
| 452 |
+
<td><br>TODO</td>
|
| 453 |
+
<td><br>TODO</td>
|
| 454 |
+
<td><br>TODO</td>
|
| 455 |
+
<td><br>TODO</td>
|
| 456 |
+
<td><br>TODO</td>
|
| 457 |
+
<td><br>TODO</td>
|
| 458 |
+
</tr>
|
| 459 |
+
<tr>
|
| 460 |
+
<td><br>Recall</td>
|
| 461 |
+
<td><br>TODO</td>
|
| 462 |
+
<td><br>TODO</td>
|
| 463 |
+
<td><br>TODO</td>
|
| 464 |
+
<td><br>TODO</td>
|
| 465 |
+
<td><br>TODO</td>
|
| 466 |
+
<td><br>TODO</td>
|
| 467 |
+
</tr>
|
| 468 |
+
<tr>
|
| 469 |
+
<td>F1</td>
|
| 470 |
+
<td><br>TODO</td>
|
| 471 |
+
<td><br>TODO</td>
|
| 472 |
+
<td><br>TODO</td>
|
| 473 |
+
<td><br>TODO</td>
|
| 474 |
+
<td><br>TODO</td>
|
| 475 |
+
<td><br>TODO</td>
|
| 476 |
+
</tr>
|
| 477 |
+
<tr>
|
| 478 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 479 |
+
<td><br>Precision</td>
|
| 480 |
+
<td><br>TODO</td>
|
| 481 |
+
<td><br>TODO</td>
|
| 482 |
+
<td><br>TODO</td>
|
| 483 |
+
<td><br>TODO</td>
|
| 484 |
+
<td><br>TODO</td>
|
| 485 |
+
<td><br>TODO</td>
|
| 486 |
+
</tr>
|
| 487 |
+
<tr>
|
| 488 |
+
<td><br>Recall</td>
|
| 489 |
+
<td><br>TODO</td>
|
| 490 |
+
<td><br>TODO</td>
|
| 491 |
+
<td><br>TODO</td>
|
| 492 |
+
<td><br>TODO</td>
|
| 493 |
+
<td><br>TODO</td>
|
| 494 |
+
<td><br>TODO</td>
|
| 495 |
+
</tr>
|
| 496 |
+
<tr>
|
| 497 |
+
<td>F1</td>
|
| 498 |
+
<td><br>TODO</td>
|
| 499 |
+
<td><br>TODO</td>
|
| 500 |
+
<td><br>TODO</td>
|
| 501 |
+
<td><br>TODO</td>
|
| 502 |
+
<td><br>TODO</td>
|
| 503 |
+
<td><br>TODO</td>
|
| 504 |
+
</tr>
|
| 505 |
+
<tr>
|
| 506 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 507 |
+
<td><br>Precision</td>
|
| 508 |
+
<td><br>0.964</td>
|
| 509 |
+
<td><br>0.922</td>
|
| 510 |
+
<td><br>0.904</td>
|
| 511 |
+
<td><br>0.856</td>
|
| 512 |
+
<td><br>0.981</td>
|
| 513 |
+
<td><br>0.961</td>
|
| 514 |
+
</tr>
|
| 515 |
+
<tr>
|
| 516 |
+
<td><br>Recall</td>
|
| 517 |
+
<td><br>0.975</td>
|
| 518 |
+
<td><br>0.917</td>
|
| 519 |
+
<td><br>0.904</td>
|
| 520 |
+
<td><br>0.872</td>
|
| 521 |
+
<td><br>0.976</td>
|
| 522 |
+
<td><br>0.961</td>
|
| 523 |
+
</tr>
|
| 524 |
+
<tr>
|
| 525 |
+
<td>F1</td>
|
| 526 |
+
<td><br><b>0.969</b></td>
|
| 527 |
+
<td><br><b>0.919</b></td>
|
| 528 |
+
<td><br><b>0.904</b></td>
|
| 529 |
+
<td><br><b>0.864</b></td>
|
| 530 |
+
<td><br><b>0.978</b></td>
|
| 531 |
+
<td><br><b>0.961</b></td>
|
| 532 |
+
</tr>
|
| 533 |
+
</tbody>
|
| 534 |
+
</table>
|
| 535 |
+
</details>
|
| 536 |
+
|
| 537 |
+
|
| 538 |
+
### multinerd
|
| 539 |
+
|
| 540 |
+
For space reasons, we show only the F1 of the different models. You can see the full results below the table.
|
| 541 |
+
|
| 542 |
+
<table>
|
| 543 |
+
<thead>
|
| 544 |
+
<tr>
|
| 545 |
+
<th><br>Model</th>
|
| 546 |
+
<th><br>PER</th>
|
| 547 |
+
<th><br>LOC</th>
|
| 548 |
+
<th><br>ORG</th>
|
| 549 |
+
<th><br>MISC</th>
|
| 550 |
+
</tr>
|
| 551 |
+
</thead>
|
| 552 |
+
<tbody>
|
| 553 |
+
<tr>
|
| 554 |
+
<td rowspan="1"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 555 |
+
<td><br>0.962</td>
|
| 556 |
+
<td><br>0.934</td>
|
| 557 |
+
<td><br>0.888</td>
|
| 558 |
+
<td><br>0.419</td>
|
| 559 |
+
</tr>
|
| 560 |
+
<tr>
|
| 561 |
+
<td rowspan="1"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 562 |
+
<td><br>0.972</td>
|
| 563 |
+
<td><br>0.938</td>
|
| 564 |
+
<td><br>0.884</td>
|
| 565 |
+
<td><br>0.430</td>
|
| 566 |
+
</tr>
|
| 567 |
+
<tr>
|
| 568 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 569 |
+
<td><br>0.985</td>
|
| 570 |
+
<td><br>0.973</td>
|
| 571 |
+
<td><br>0.938</td>
|
| 572 |
+
<td><br>0.770</td>
|
| 573 |
+
</tr>
|
| 574 |
+
<tr>
|
| 575 |
+
<td rowspan="1"><br>NERmembert2-4entities (this model) (111M)</td>
|
| 576 |
+
<td><br>TODO</td>
|
| 577 |
+
<td><br>TODO</td>
|
| 578 |
+
<td><br>TODO</td>
|
| 579 |
+
<td><br>TODO</td>
|
| 580 |
+
</tr>
|
| 581 |
+
<tr>
|
| 582 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 583 |
+
<td><br>TODO</td>
|
| 584 |
+
<td><br>TODO</td>
|
| 585 |
+
<td><br>TODO</td>
|
| 586 |
+
<td><br>TODO</td>
|
| 587 |
+
</tr>
|
| 588 |
+
<tr>
|
| 589 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 590 |
+
<td><br><b>0.987</b></td>
|
| 591 |
+
<td><br>0.976</td>
|
| 592 |
+
<td><br>0.948</td>
|
| 593 |
+
<td><br><b>0.790</b></td>
|
| 594 |
+
</tr>
|
| 595 |
+
</tbody>
|
| 596 |
+
</table>
|
| 597 |
+
|
| 598 |
+
<details>
|
| 599 |
+
<summary>Full results</summary>
|
| 600 |
+
<table>
|
| 601 |
+
<thead>
|
| 602 |
+
<tr>
|
| 603 |
+
<th><br>Model</th>
|
| 604 |
+
<th><br>Metrics</th>
|
| 605 |
+
<th><br>PER</th>
|
| 606 |
+
<th><br>LOC</th>
|
| 607 |
+
<th><br>ORG</th>
|
| 608 |
+
<th><br>MISC</th>
|
| 609 |
+
<th><br>O</th>
|
| 610 |
+
<th><br>Overall</th>
|
| 611 |
+
</tr>
|
| 612 |
+
</thead>
|
| 613 |
+
<tbody>
|
| 614 |
+
<tr>
|
| 615 |
+
<td rowspan="3"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 616 |
+
<td><br>Precision</td>
|
| 617 |
+
<td><br>0.931</td>
|
| 618 |
+
<td><br>0.893</td>
|
| 619 |
+
<td><br>0.827</td>
|
| 620 |
+
<td><br>0.725</td>
|
| 621 |
+
<td><br>0.979</td>
|
| 622 |
+
<td><br>0.966</td>
|
| 623 |
+
</tr>
|
| 624 |
+
<tr>
|
| 625 |
+
<td><br>Recall</td>
|
| 626 |
+
<td><br>0.994</td>
|
| 627 |
+
<td><br>0.980</td>
|
| 628 |
+
<td><br>0.959</td>
|
| 629 |
+
<td><br>0.295</td>
|
| 630 |
+
<td><br>0.990</td>
|
| 631 |
+
<td><br>0.966</td>
|
| 632 |
+
</tr>
|
| 633 |
+
<tr>
|
| 634 |
+
<td>F1</td>
|
| 635 |
+
<td><br>0.962</td>
|
| 636 |
+
<td><br>0.934</td>
|
| 637 |
+
<td><br>0.888</td>
|
| 638 |
+
<td><br>0.419</td>
|
| 639 |
+
<td><br>0.984</td>
|
| 640 |
+
<td><br>0.966</td>
|
| 641 |
+
</tr>
|
| 642 |
+
<tr>
|
| 643 |
+
<td rowspan="3"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 644 |
+
<td><br>Precision</td>
|
| 645 |
+
<td><br>0.954</td>
|
| 646 |
+
<td><br>0.908</td>
|
| 647 |
+
<td><br>0.817</td>
|
| 648 |
+
<td><br>0.705</td>
|
| 649 |
+
<td><br>0.977</td>
|
| 650 |
+
<td><br>0.967</td>
|
| 651 |
+
</tr>
|
| 652 |
+
<tr>
|
| 653 |
+
<td><br>Recall</td>
|
| 654 |
+
<td><br>0.991</td>
|
| 655 |
+
<td><br>0.969</td>
|
| 656 |
+
<td><br>0.963</td>
|
| 657 |
+
<td><br>0.310</td>
|
| 658 |
+
<td><br>0.990</td>
|
| 659 |
+
<td><br>0.967</td>
|
| 660 |
+
</tr>
|
| 661 |
+
<tr>
|
| 662 |
+
<td>F1</td>
|
| 663 |
+
<td><br>0.972</td>
|
| 664 |
+
<td><br>0.938</td>
|
| 665 |
+
<td><br>0.884</td>
|
| 666 |
+
<td><br>0.430</td>
|
| 667 |
+
<td><br>0.984</td>
|
| 668 |
+
<td><br>0.967</td>
|
| 669 |
+
</tr>
|
| 670 |
+
<tr>
|
| 671 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 672 |
+
<td><br>Precision</td>
|
| 673 |
+
<td><br>0.976</td>
|
| 674 |
+
<td><br>0.961</td>
|
| 675 |
+
<td><br>0.911</td>
|
| 676 |
+
<td><br>0.829</td>
|
| 677 |
+
<td><br>0.991</td>
|
| 678 |
+
<td><br>0.983</td>
|
| 679 |
+
</tr>
|
| 680 |
+
<tr>
|
| 681 |
+
<td><br>Recall</td>
|
| 682 |
+
<td><br>0.994</td>
|
| 683 |
+
<td><br>0.985</td>
|
| 684 |
+
<td><br>0.967</td>
|
| 685 |
+
<td><br>0.719</td>
|
| 686 |
+
<td><br>0.993</td>
|
| 687 |
+
<td><br>0.983</td>
|
| 688 |
+
</tr>
|
| 689 |
+
<tr>
|
| 690 |
+
<td>F1</td>
|
| 691 |
+
<td><br>0.985</td>
|
| 692 |
+
<td><br>0.973</td>
|
| 693 |
+
<td><br>0.938</td>
|
| 694 |
+
<td><br>0.770</td>
|
| 695 |
+
<td><br>0.992</td>
|
| 696 |
+
<td><br>0.983</td>
|
| 697 |
+
</tr>
|
| 698 |
+
<tr>
|
| 699 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert2-4entities">NERmembert2-4entities (this model) (111M)</a></td>
|
| 700 |
+
<td><br>Precision</td>
|
| 701 |
+
<td><br>TODO</td>
|
| 702 |
+
<td><br>TODO</td>
|
| 703 |
+
<td><br>TODO</td>
|
| 704 |
+
<td><br>TODO</td>
|
| 705 |
+
<td><br>TODO</td>
|
| 706 |
+
<td><br>TODO</td>
|
| 707 |
+
</tr>
|
| 708 |
+
<tr>
|
| 709 |
+
<td><br>Recall</td>
|
| 710 |
+
<td><br>TODO</td>
|
| 711 |
+
<td><br>TODO</td>
|
| 712 |
+
<td><br>TODO</td>
|
| 713 |
+
<td><br>TODO</td>
|
| 714 |
+
<td><br>TODO</td>
|
| 715 |
+
<td><br>TODO</td>
|
| 716 |
+
</tr>
|
| 717 |
+
<tr>
|
| 718 |
+
<td>F1</td>
|
| 719 |
+
<td><br>TODO</td>
|
| 720 |
+
<td><br>TODO</td>
|
| 721 |
+
<td><br>TODO</td>
|
| 722 |
+
<td><br>TODO</td>
|
| 723 |
+
<td><br>TODO</td>
|
| 724 |
+
<td><br>TODO</td>
|
| 725 |
+
</tr>
|
| 726 |
+
<tr>
|
| 727 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 728 |
+
<td><br>Precision</td>
|
| 729 |
+
<td><br>TODO</td>
|
| 730 |
+
<td><br>TODO</td>
|
| 731 |
+
<td><br>TODO</td>
|
| 732 |
+
<td><br>TODO</td>
|
| 733 |
+
<td><br>TODO</td>
|
| 734 |
+
<td><br>TODO</td>
|
| 735 |
+
</tr>
|
| 736 |
+
<tr>
|
| 737 |
+
<td><br>Recall</td>
|
| 738 |
+
<td><br>TODO</td>
|
| 739 |
+
<td><br>TODO</td>
|
| 740 |
+
<td><br>TODO</td>
|
| 741 |
+
<td><br>TODO</td>
|
| 742 |
+
<td><br>TODO</td>
|
| 743 |
+
<td><br>TODO</td>
|
| 744 |
+
</tr>
|
| 745 |
+
<tr>
|
| 746 |
+
<td>F1</td>
|
| 747 |
+
<td><br>TODO</td>
|
| 748 |
+
<td><br>TODO</td>
|
| 749 |
+
<td><br>TODO</td>
|
| 750 |
+
<td><br>TODO</td>
|
| 751 |
+
<td><br>TODO</td>
|
| 752 |
+
<td><br>TODO</td>
|
| 753 |
+
</tr>
|
| 754 |
+
<tr>
|
| 755 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 756 |
+
<td><br>Precision</td>
|
| 757 |
+
<td><br>0.979</td>
|
| 758 |
+
<td><br>0.967</td>
|
| 759 |
+
<td><br>0.922</td>
|
| 760 |
+
<td><br>0.852</td>
|
| 761 |
+
<td><br>0.991</td>
|
| 762 |
+
<td><br>0.985</td>
|
| 763 |
+
</tr>
|
| 764 |
+
<tr>
|
| 765 |
+
<td><br>Recall</td>
|
| 766 |
+
<td><br>0.996</td>
|
| 767 |
+
<td><br>0.986</td>
|
| 768 |
+
<td><br>0.974</td>
|
| 769 |
+
<td><br>0.736</td>
|
| 770 |
+
<td><br>0.994</td>
|
| 771 |
+
<td><br>0.985</td>
|
| 772 |
+
</tr>
|
| 773 |
+
<tr>
|
| 774 |
+
<td>F1</td>
|
| 775 |
+
<td><br><b>0.987</b></td>
|
| 776 |
+
<td><br>0.976</td>
|
| 777 |
+
<td><br>0.948</td>
|
| 778 |
+
<td><br><b>0.790</b></td>
|
| 779 |
+
<td><br>0.993</td>
|
| 780 |
+
<td><br>0.985</td>
|
| 781 |
+
</tr>
|
| 782 |
+
</tbody>
|
| 783 |
+
</table>
|
| 784 |
+
</details>
|
| 785 |
+
|
| 786 |
+
### wikiner
|
| 787 |
+
|
| 788 |
+
For space reasons, we show only the F1 of the different models. You can see the full results below the table.
|
| 789 |
+
|
| 790 |
+
<table>
|
| 791 |
+
<thead>
|
| 792 |
+
<tr>
|
| 793 |
+
<th><br>Model</th>
|
| 794 |
+
<th><br>PER</th>
|
| 795 |
+
<th><br>LOC</th>
|
| 796 |
+
<th><br>ORG</th>
|
| 797 |
+
<th><br>MISC</th>
|
| 798 |
+
</tr>
|
| 799 |
+
</thead>
|
| 800 |
+
<tbody>
|
| 801 |
+
<tr>
|
| 802 |
+
<td rowspan="1"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 803 |
+
<td><br><b>0.986</b></td>
|
| 804 |
+
<td><br><b>0.966</b></td>
|
| 805 |
+
<td><br><b>0.938</b></td>
|
| 806 |
+
<td><br><b>0.938</b></td>
|
| 807 |
+
</tr>
|
| 808 |
+
<tr>
|
| 809 |
+
<td rowspan="1"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 810 |
+
<td><br>0.983</td>
|
| 811 |
+
<td><br>0.964</td>
|
| 812 |
+
<td><br>0.925</td>
|
| 813 |
+
<td><br>0.926</td>
|
| 814 |
+
</tr>
|
| 815 |
+
<tr>
|
| 816 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 817 |
+
<td><br>0.970</td>
|
| 818 |
+
<td><br>0.945</td>
|
| 819 |
+
<td><br>0.876</td>
|
| 820 |
+
<td><br>0.872</td>
|
| 821 |
+
</tr>
|
| 822 |
+
<tr>
|
| 823 |
+
<td rowspan="1"><br>NERmembert2-4entities (this model) (111M)</td>
|
| 824 |
+
<td><br>TODO</td>
|
| 825 |
+
<td><br>TODO</td>
|
| 826 |
+
<td><br>TODO</td>
|
| 827 |
+
<td><br>TODO</td>
|
| 828 |
+
</tr>
|
| 829 |
+
<tr>
|
| 830 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 831 |
+
<td><br>TODO</td>
|
| 832 |
+
<td><br>TODO</td>
|
| 833 |
+
<td><br>TODO</td>
|
| 834 |
+
<td><br>TODO</td>
|
| 835 |
+
</tr>
|
| 836 |
+
<tr>
|
| 837 |
+
<td rowspan="1"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 838 |
+
<td><br>0.975</td>
|
| 839 |
+
<td><br>0.953</td>
|
| 840 |
+
<td><br>0.896</td>
|
| 841 |
+
<td><br>0.893</td>
|
| 842 |
+
</tr>
|
| 843 |
+
</tbody>
|
| 844 |
+
</table>
|
| 845 |
+
|
| 846 |
+
<details>
|
| 847 |
+
<summary>Full results</summary>
|
| 848 |
+
<table>
|
| 849 |
+
<thead>
|
| 850 |
+
<tr>
|
| 851 |
+
<th><br>Model</th>
|
| 852 |
+
<th><br>Metrics</th>
|
| 853 |
+
<th><br>PER</th>
|
| 854 |
+
<th><br>LOC</th>
|
| 855 |
+
<th><br>ORG</th>
|
| 856 |
+
<th><br>MISC</th>
|
| 857 |
+
<th><br>O</th>
|
| 858 |
+
<th><br>Overall</th>
|
| 859 |
+
</tr>
|
| 860 |
+
</thead>
|
| 861 |
+
<tbody>
|
| 862 |
+
<tr>
|
| 863 |
+
<td rowspan="3"><br><a href="https://hf.co/Jean-Baptiste/camembert-ner">Jean-Baptiste/camembert-ner (110M)</a></td>
|
| 864 |
+
<td><br>Precision</td>
|
| 865 |
+
<td><br>0.986</td>
|
| 866 |
+
<td><br>0.962</td>
|
| 867 |
+
<td><br>0.925</td>
|
| 868 |
+
<td><br>0.943</td>
|
| 869 |
+
<td><br>0.998</td>
|
| 870 |
+
<td><br>0.992</td>
|
| 871 |
+
</tr>
|
| 872 |
+
<tr>
|
| 873 |
+
<td><br>Recall</td>
|
| 874 |
+
<td><br>0.987</td>
|
| 875 |
+
<td><br>0.969</td>
|
| 876 |
+
<td><br>0.951</td>
|
| 877 |
+
<td><br>0.933</td>
|
| 878 |
+
<td><br>0.997</td>
|
| 879 |
+
<td><br>0.992</td>
|
| 880 |
+
</tr>
|
| 881 |
+
<tr>
|
| 882 |
+
<td>F1</td>
|
| 883 |
+
<td><br><b>0.986</b></td>
|
| 884 |
+
<td><br><b>0.966</b></td>
|
| 885 |
+
<td><br><b>0.938</b></td>
|
| 886 |
+
<td><br><b>0.938</b></td>
|
| 887 |
+
<td><br><b>0.998</b></td>
|
| 888 |
+
<td><br><b>0.992</b></td>
|
| 889 |
+
</tr>
|
| 890 |
+
<tr>
|
| 891 |
+
<td rowspan="3"><br><a href="https://hf.co/cmarkea/distilcamembert-base-ner">cmarkea/distilcamembert-base-ner (67.5M)</a></td>
|
| 892 |
+
<td><br>Precision</td>
|
| 893 |
+
<td><br>0.982</td>
|
| 894 |
+
<td><br>0.964</td>
|
| 895 |
+
<td><br>0.910</td>
|
| 896 |
+
<td><br>0.942</td>
|
| 897 |
+
<td><br>0.997</td>
|
| 898 |
+
<td><br>0.991</td>
|
| 899 |
+
</tr>
|
| 900 |
+
<tr>
|
| 901 |
+
<td><br>Recall</td>
|
| 902 |
+
<td><br>0.985</td>
|
| 903 |
+
<td><br>0.963</td>
|
| 904 |
+
<td><br>0.940</td>
|
| 905 |
+
<td><br>0.910</td>
|
| 906 |
+
<td><br>0.998</td>
|
| 907 |
+
<td><br>0.991</td>
|
| 908 |
+
</tr>
|
| 909 |
+
<tr>
|
| 910 |
+
<td>F1</td>
|
| 911 |
+
<td><br>0.983</td>
|
| 912 |
+
<td><br>0.964</td>
|
| 913 |
+
<td><br>0.925</td>
|
| 914 |
+
<td><br>0.926</td>
|
| 915 |
+
<td><br>0.997</td>
|
| 916 |
+
<td><br>0.991</td>
|
| 917 |
+
</tr>
|
| 918 |
+
<tr>
|
| 919 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-base-4entities">NERmembert-base-4entities</a></td>
|
| 920 |
+
<td><br>Precision</td>
|
| 921 |
+
<td><br>0.970</td>
|
| 922 |
+
<td><br>0.944</td>
|
| 923 |
+
<td><br>0.872</td>
|
| 924 |
+
<td><br>0.878</td>
|
| 925 |
+
<td><br>0.996</td>
|
| 926 |
+
<td><br>0.986</td>
|
| 927 |
+
</tr>
|
| 928 |
+
<tr>
|
| 929 |
+
<td><br>Recall</td>
|
| 930 |
+
<td><br>0.969</td>
|
| 931 |
+
<td><br>0.947</td>
|
| 932 |
+
<td><br>0.880</td>
|
| 933 |
+
<td><br>0.866</td>
|
| 934 |
+
<td><br>0.996</td>
|
| 935 |
+
<td><br>0.986</td>
|
| 936 |
+
</tr>
|
| 937 |
+
<tr>
|
| 938 |
+
<td>F1</td>
|
| 939 |
+
<td><br>0.970</td>
|
| 940 |
+
<td><br>0.945</td>
|
| 941 |
+
<td><br>0.876</td>
|
| 942 |
+
<td><br>0.872</td>
|
| 943 |
+
<td><br>0.996</td>
|
| 944 |
+
<td><br>0.986</td>
|
| 945 |
+
</tr>
|
| 946 |
+
<tr>
|
| 947 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert2-4entities">NERmembert2-4entities (this model) (111M)</a></td>
|
| 948 |
+
<td><br>Precision</td>
|
| 949 |
+
<td><br>TODO</td>
|
| 950 |
+
<td><br>TODO</td>
|
| 951 |
+
<td><br>TODO</td>
|
| 952 |
+
<td><br>TODO</td>
|
| 953 |
+
<td><br>TODO</td>
|
| 954 |
+
<td><br>TODO</td>
|
| 955 |
+
</tr>
|
| 956 |
+
<tr>
|
| 957 |
+
<td><br>Recall</td>
|
| 958 |
+
<td><br>TODO</td>
|
| 959 |
+
<td><br>TODO</td>
|
| 960 |
+
<td><br>TODO</td>
|
| 961 |
+
<td><br>TODO</td>
|
| 962 |
+
<td><br>TODO</td>
|
| 963 |
+
<td><br>TODO</td>
|
| 964 |
+
</tr>
|
| 965 |
+
<tr>
|
| 966 |
+
<td>F1</td>
|
| 967 |
+
<td><br>TODO</td>
|
| 968 |
+
<td><br>TODO</td>
|
| 969 |
+
<td><br>TODO</td>
|
| 970 |
+
<td><br>TODO</td>
|
| 971 |
+
<td><br>TODO</td>
|
| 972 |
+
<td><br>TODO</td>
|
| 973 |
+
</tr>
|
| 974 |
+
<tr>
|
| 975 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmemberta-4entities">NERmemberta-4entities (111M)</a></td>
|
| 976 |
+
<td><br>Precision</td>
|
| 977 |
+
<td><br>TODO</td>
|
| 978 |
+
<td><br>TODO</td>
|
| 979 |
+
<td><br>TODO</td>
|
| 980 |
+
<td><br>TODO</td>
|
| 981 |
+
<td><br>TODO</td>
|
| 982 |
+
<td><br>TODO</td>
|
| 983 |
+
</tr>
|
| 984 |
+
<tr>
|
| 985 |
+
<td><br>Recall</td>
|
| 986 |
+
<td><br>TODO</td>
|
| 987 |
+
<td><br>TODO</td>
|
| 988 |
+
<td><br>TODO</td>
|
| 989 |
+
<td><br>TODO</td>
|
| 990 |
+
<td><br>TODO</td>
|
| 991 |
+
<td><br>TODO</td>
|
| 992 |
+
</tr>
|
| 993 |
+
<tr>
|
| 994 |
+
<td>F1</td>
|
| 995 |
+
<td><br>TODO</td>
|
| 996 |
+
<td><br>TODO</td>
|
| 997 |
+
<td><br>TODO</td>
|
| 998 |
+
<td><br>TODO</td>
|
| 999 |
+
<td><br>TODO</td>
|
| 1000 |
+
<td><br>TODO</td>
|
| 1001 |
+
</tr>
|
| 1002 |
+
<tr>
|
| 1003 |
+
<td rowspan="3"><br><a href="https://hf.co/CATIE-AQ/NERmembert-large-4entities">NERmembert-large-4entities (336M)</a></td>
|
| 1004 |
+
<td><br>Precision</td>
|
| 1005 |
+
<td><br>0.975</td>
|
| 1006 |
+
<td><br>0.957</td>
|
| 1007 |
+
<td><br>0.872</td>
|
| 1008 |
+
<td><br>0.901</td>
|
| 1009 |
+
<td><br>0.997</td>
|
| 1010 |
+
<td><br>0.989</td>
|
| 1011 |
+
</tr>
|
| 1012 |
+
<tr>
|
| 1013 |
+
<td><br>Recall</td>
|
| 1014 |
+
<td><br>0.975</td>
|
| 1015 |
+
<td><br>0.949</td>
|
| 1016 |
+
<td><br>0.922</td>
|
| 1017 |
+
<td><br>0.884</td>
|
| 1018 |
+
<td><br>0.997</td>
|
| 1019 |
+
<td><br>0.989</td>
|
| 1020 |
+
</tr>
|
| 1021 |
+
<tr>
|
| 1022 |
+
<td>F1</td>
|
| 1023 |
+
<td><br>0.975</td>
|
| 1024 |
+
<td><br>0.953</td>
|
| 1025 |
+
<td><br>0.896</td>
|
| 1026 |
+
<td><br>0.893</td>
|
| 1027 |
+
<td><br>0.997</td>
|
| 1028 |
+
<td><br>0.989</td>
|
| 1029 |
+
</tr>
|
| 1030 |
+
</tbody>
|
| 1031 |
+
</table>
|
| 1032 |
+
</details>
|
| 1033 |
+
|
| 1034 |
+
|
| 1035 |
+
## Usage
|
| 1036 |
+
### Code
|
| 1037 |
+
|
| 1038 |
+
```python
|
| 1039 |
+
from transformers import pipeline
|
| 1040 |
+
|
| 1041 |
+
ner = pipeline('token-classification', model='CATIE-AQ/NERmembert2-4entities', tokenizer='CATIE-AQ/NERmembert2-4entities', aggregation_strategy="simple")
|
| 1042 |
+
|
| 1043 |
+
result = ner(
|
| 1044 |
+
"Le dévoilement du logo officiel des JO s'est déroulé le 21 octobre 2019 au Grand Rex. Ce nouvel emblème et cette nouvelle typographie ont été conçus par le designer Sylvain Boyer avec les agences Royalties & Ecobranding. Rond, il rassemble trois symboles : une médaille d'or, la flamme olympique et Marianne, symbolisée par un visage de femme mais privée de son bonnet phrygien caractéristique. La typographie dessinée fait référence à l'Art déco, mouvement artistique des années 1920, décennie pendant laquelle ont eu lieu pour la dernière fois les Jeux olympiques à Paris en 1924. Pour la première fois, ce logo sera unique pour les Jeux olympiques et les Jeux paralympiques."
|
| 1045 |
+
)
|
| 1046 |
+
|
| 1047 |
+
print(result)
|
| 1048 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1049 |
|
|
|
|
| 1050 |
|
| 1051 |
+
### Try it through Space
|
| 1052 |
+
A Space has been created to test the model. It is available [here](https://huggingface.co/spaces/CATIE-AQ/NERmembert).
|
| 1053 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1054 |
|
| 1055 |
+
## Environmental Impact
|
| 1056 |
|
| 1057 |
+
*Carbon emissions were estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). The hardware, runtime, cloud provider, and compute region were utilized to estimate the carbon impact.*
|
| 1058 |
|
| 1059 |
+
- **Hardware Type:** A100 PCIe 40/80GB
|
| 1060 |
+
- **Hours used:** 1h51min
|
| 1061 |
+
- **Cloud Provider:** Private Infrastructure
|
| 1062 |
+
- **Carbon Efficiency (kg/kWh):** 0.055 (estimated from [electricitymaps](https://app.electricitymaps.com/zone/FR) for the day of November 21, 2024.)
|
| 1063 |
+
- **Carbon Emitted** *(Power consumption x Time x Carbon produced based on location of power grid)*: 0.0255 kg eq. CO2
|
| 1064 |
|
|
|
|
| 1065 |
|
|
|
|
| 1066 |
|
| 1067 |
+
## Citations
|
| 1068 |
+
|
| 1069 |
+
### NERmemBERT2-4entities
|
| 1070 |
+
```
|
| 1071 |
+
@misc {NERmemberta2024,
|
| 1072 |
+
author = { {BOURDOIS, Loïck} },
|
| 1073 |
+
organization = { {Centre Aquitain des Technologies de l'Information et Electroniques} },
|
| 1074 |
+
title = { NERmemberta-base-3entities },
|
| 1075 |
+
year = 2024,
|
| 1076 |
+
url = { https://huggingface.co/CATIE-AQ/NERmemberta-base-3entities },
|
| 1077 |
+
doi = { 10.57967/hf/1752 },
|
| 1078 |
+
publisher = { Hugging Face }
|
| 1079 |
+
}
|
| 1080 |
+
```
|
| 1081 |
+
|
| 1082 |
+
### NERmemBERT
|
| 1083 |
+
```
|
| 1084 |
+
@misc {NERmembert2024,
|
| 1085 |
+
author = { {BOURDOIS, Loïck} },
|
| 1086 |
+
organization = { {Centre Aquitain des Technologies de l'Information et Electroniques} },
|
| 1087 |
+
title = { NERmembert-base-3entities },
|
| 1088 |
+
year = 2024,
|
| 1089 |
+
url = { https://huggingface.co/CATIE-AQ/NERmembert-base-4entities },
|
| 1090 |
+
doi = { 10.57967/hf/1752 },
|
| 1091 |
+
publisher = { Hugging Face }
|
| 1092 |
+
}
|
| 1093 |
+
```
|
| 1094 |
+
|
| 1095 |
+
### CamemBERT
|
| 1096 |
+
```
|
| 1097 |
+
@inproceedings{martin2020camembert,
|
| 1098 |
+
title={CamemBERT: a Tasty French Language Model},
|
| 1099 |
+
author={Martin, Louis and Muller, Benjamin and Su{\'a}rez, Pedro Javier Ortiz and Dupont, Yoann and Romary, Laurent and de la Clergerie, {\'E}ric Villemonte and Seddah, Djam{\'e} and Sagot, Beno{\^\i}t},
|
| 1100 |
+
booktitle={Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics},
|
| 1101 |
+
year={2020}}
|
| 1102 |
+
```
|
| 1103 |
|
| 1104 |
+
### CamemBERT 2.0
|
| 1105 |
+
```
|
| 1106 |
+
@misc{antoun2024camembert20smarterfrench,
|
| 1107 |
+
title={CamemBERT 2.0: A Smarter French Language Model Aged to Perfection},
|
| 1108 |
+
author={Wissam Antoun and Francis Kulumba and Rian Touchent and Éric de la Clergerie and Benoît Sagot and Djamé Seddah},
|
| 1109 |
+
year={2024},
|
| 1110 |
+
eprint={2411.08868},
|
| 1111 |
+
archivePrefix={arXiv},
|
| 1112 |
+
primaryClass={cs.CL},
|
| 1113 |
+
url={https://arxiv.org/abs/2411.08868},
|
| 1114 |
+
}
|
| 1115 |
+
```
|
| 1116 |
|
| 1117 |
+
### multiconer
|
| 1118 |
+
```
|
| 1119 |
+
@inproceedings{multiconer2-report,
|
| 1120 |
+
title={{SemEval-2023 Task 2: Fine-grained Multilingual Named Entity Recognition (MultiCoNER 2)}},
|
| 1121 |
+
author={Fetahu, Besnik and Kar, Sudipta and Chen, Zhiyu and Rokhlenko, Oleg and Malmasi, Shervin},
|
| 1122 |
+
booktitle={Proceedings of the 17th International Workshop on Semantic Evaluation (SemEval-2023)},
|
| 1123 |
+
year={2023},
|
| 1124 |
+
publisher={Association for Computational Linguistics}}
|
| 1125 |
|
| 1126 |
+
@article{multiconer2-data,
|
| 1127 |
+
title={{MultiCoNER v2: a Large Multilingual dataset for Fine-grained and Noisy Named Entity Recognition}},
|
| 1128 |
+
author={Fetahu, Besnik and Chen, Zhiyu and Kar, Sudipta and Rokhlenko, Oleg and Malmasi, Shervin},
|
| 1129 |
+
year={2023}}
|
| 1130 |
+
```
|
|
|
|
|
|
|
|
|
|
| 1131 |
|
| 1132 |
+
### multinerd
|
| 1133 |
+
```
|
| 1134 |
+
@inproceedings{tedeschi-navigli-2022-multinerd,
|
| 1135 |
+
title = "{M}ulti{NERD}: A Multilingual, Multi-Genre and Fine-Grained Dataset for Named Entity Recognition (and Disambiguation)",
|
| 1136 |
+
author = "Tedeschi, Simone and Navigli, Roberto",
|
| 1137 |
+
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2022",
|
| 1138 |
+
month = jul,
|
| 1139 |
+
year = "2022",
|
| 1140 |
+
address = "Seattle, United States",
|
| 1141 |
+
publisher = "Association for Computational Linguistics",
|
| 1142 |
+
url = "https://aclanthology.org/2022.findings-naacl.60",
|
| 1143 |
+
doi = "10.18653/v1/2022.findings-naacl.60",
|
| 1144 |
+
pages = "801--812"}
|
| 1145 |
+
```
|
| 1146 |
|
| 1147 |
+
### pii-masking-200k
|
| 1148 |
+
```
|
| 1149 |
+
@misc {ai4privacy_2023,
|
| 1150 |
+
author = { {ai4Privacy} },
|
| 1151 |
+
title = { pii-masking-200k (Revision 1d4c0a1) },
|
| 1152 |
+
year = 2023,
|
| 1153 |
+
url = { https://huggingface.co/datasets/ai4privacy/pii-masking-200k },
|
| 1154 |
+
doi = { 10.57967/hf/1532 },
|
| 1155 |
+
publisher = { Hugging Face }}
|
| 1156 |
+
```
|
| 1157 |
|
| 1158 |
+
### wikiner
|
| 1159 |
+
```
|
| 1160 |
+
@article{NOTHMAN2013151,
|
| 1161 |
+
title = {Learning multilingual named entity recognition from Wikipedia},
|
| 1162 |
+
journal = {Artificial Intelligence},
|
| 1163 |
+
volume = {194},
|
| 1164 |
+
pages = {151-175},
|
| 1165 |
+
year = {2013},
|
| 1166 |
+
note = {Artificial Intelligence, Wikipedia and Semi-Structured Resources},
|
| 1167 |
+
issn = {0004-3702},
|
| 1168 |
+
doi = {https://doi.org/10.1016/j.artint.2012.03.006},
|
| 1169 |
+
url = {https://www.sciencedirect.com/science/article/pii/S0004370212000276},
|
| 1170 |
+
author = {Joel Nothman and Nicky Ringland and Will Radford and Tara Murphy and James R. Curran}}
|
| 1171 |
+
```
|
| 1172 |
|
| 1173 |
+
### frenchNER_4entities
|
| 1174 |
+
```
|
| 1175 |
+
@misc {frenchNER2024,
|
| 1176 |
+
author = { {BOURDOIS, Loïck} },
|
| 1177 |
+
organization = { {Centre Aquitain des Technologies de l'Information et Electroniques} },
|
| 1178 |
+
title = { frenchNER_4entities },
|
| 1179 |
+
year = 2024,
|
| 1180 |
+
url = { https://huggingface.co/CATIE-AQ/frenchNER_4entities },
|
| 1181 |
+
doi = { 10.57967/hf/1751 },
|
| 1182 |
+
publisher = { Hugging Face }
|
| 1183 |
+
}
|
| 1184 |
+
```
|
| 1185 |
|
| 1186 |
+
## License
|
| 1187 |
+
MIT
|
|
|
|
|
|