Commit
•
93a9a2f
1
Parent(s):
44e2be3
Update README.md
Browse files
README.md
CHANGED
@@ -87,6 +87,16 @@ It contains the following tasks and their related datasets:
|
|
87 |
|
88 |
Here are the train/dev/test splits of the datasets:
|
89 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
90 |
| Task | NER (F1) | POS (F1) | STS (Pearson) | TC (accuracy) | QA (ViquiQuAD) (F1/EM) | QA (XQuAD) (F1/EM) |
|
91 |
| ------------|:-------------:| -----:|:------|:-------|:------|:----|
|
92 |
| RoBERTa-base-ca-v2 | **89.80** | **99.10** | **80.00** | **83.40** | **88.00** | **71.50** |
|
@@ -95,8 +105,6 @@ Here are the train/dev/test splits of the datasets:
|
|
95 |
| XLM-RoBERTa | 87.66 | 98.89 | 75.40 | 71.68 | 85.50/70.47 | 67.10/46.42 |
|
96 |
| WikiBERT-ca | 77.66 | 97.60 | 77.18 | 73.22 | 85.45/70.75 | 65.21/36.60 |
|
97 |
|
98 |
-
### Results
|
99 |
-
|
100 |
## Intended uses & limitations
|
101 |
The model is ready-to-use only for masked language modelling to perform the Fill Mask task (try the inference API or read the next section)
|
102 |
However, the is intended to be fine-tuned on non-generative downstream tasks such as Question Answering, Text Classification or Named Entity Recognition.
|
|
|
87 |
|
88 |
Here are the train/dev/test splits of the datasets:
|
89 |
|
90 |
+
| Task (Dataset) | Total | Train | Dev | Test |
|
91 |
+
|:--|:--|:--|:--|:--|
|
92 |
+
| NER (Ancora) |13,581 | 10,628 | 1,427 | 1,526 |
|
93 |
+
| POS (Ancora)| 16,678 | 13,123 | 1,709 | 1,846 |
|
94 |
+
| STS | 3,073 | 2,073 | 500 | 500 |
|
95 |
+
| TC (TeCla) | 137,775 | 110,203 | 13,786 | 13,786|
|
96 |
+
| QA (ViquiQuAD) | 14,239 | 11,255 | 1,492 | 1,429 |
|
97 |
+
|
98 |
+
### Results
|
99 |
+
|
100 |
| Task | NER (F1) | POS (F1) | STS (Pearson) | TC (accuracy) | QA (ViquiQuAD) (F1/EM) | QA (XQuAD) (F1/EM) |
|
101 |
| ------------|:-------------:| -----:|:------|:-------|:------|:----|
|
102 |
| RoBERTa-base-ca-v2 | **89.80** | **99.10** | **80.00** | **83.40** | **88.00** | **71.50** |
|
|
|
105 |
| XLM-RoBERTa | 87.66 | 98.89 | 75.40 | 71.68 | 85.50/70.47 | 67.10/46.42 |
|
106 |
| WikiBERT-ca | 77.66 | 97.60 | 77.18 | 73.22 | 85.45/70.75 | 65.21/36.60 |
|
107 |
|
|
|
|
|
108 |
## Intended uses & limitations
|
109 |
The model is ready-to-use only for masked language modelling to perform the Fill Mask task (try the inference API or read the next section)
|
110 |
However, the is intended to be fine-tuned on non-generative downstream tasks such as Question Answering, Text Classification or Named Entity Recognition.
|