Commit ·
edfe8a8
1
Parent(s): c4538a8
updated readme
Browse files
README.md
CHANGED
|
@@ -32,11 +32,6 @@ It achieves the following results on the evaluation set:
|
|
| 32 |
- Accuracy: 0.9741
|
| 33 |
- F1: 0.8684
|
| 34 |
|
| 35 |
-
## Model description
|
| 36 |
-
|
| 37 |
-
More information needed
|
| 38 |
-
|
| 39 |
-
|
| 40 |
### Use a pipeline as a high-level helper
|
| 41 |
```python
|
| 42 |
from transformers import pipeline
|
|
@@ -51,41 +46,3 @@ from transformers import AutoTokenizer, AutoModelForTokenClassification
|
|
| 51 |
tokenizer = AutoTokenizer.from_pretrained("Azma-AI/bert-uncased-keyword-extractor")
|
| 52 |
model = AutoModelForTokenClassification.from_pretrained("Azma-AI/bert-uncased-keyword-extractor")
|
| 53 |
```
|
| 54 |
-
## Training and evaluation data
|
| 55 |
-
|
| 56 |
-
More information needed
|
| 57 |
-
|
| 58 |
-
## Training procedure
|
| 59 |
-
|
| 60 |
-
### Training hyperparameters
|
| 61 |
-
|
| 62 |
-
The following hyperparameters were used during training:
|
| 63 |
-
- learning_rate: 2e-05
|
| 64 |
-
- train_batch_size: 16
|
| 65 |
-
- eval_batch_size: 16
|
| 66 |
-
- seed: 42
|
| 67 |
-
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 68 |
-
- lr_scheduler_type: linear
|
| 69 |
-
- num_epochs: 8
|
| 70 |
-
- mixed_precision_training: Native AMP
|
| 71 |
-
|
| 72 |
-
### Training results
|
| 73 |
-
|
| 74 |
-
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | Accuracy | F1 |
|
| 75 |
-
|:-------------:|:-----:|:-----:|:---------------:|:---------:|:------:|:--------:|:------:|
|
| 76 |
-
| 0.165 | 1.0 | 1875 | 0.1202 | 0.7109 | 0.7766 | 0.9505 | 0.7423 |
|
| 77 |
-
| 0.1211 | 2.0 | 3750 | 0.1011 | 0.7801 | 0.8186 | 0.9621 | 0.7989 |
|
| 78 |
-
| 0.0847 | 3.0 | 5625 | 0.0945 | 0.8292 | 0.8044 | 0.9667 | 0.8166 |
|
| 79 |
-
| 0.0614 | 4.0 | 7500 | 0.0927 | 0.8409 | 0.8524 | 0.9711 | 0.8466 |
|
| 80 |
-
| 0.0442 | 5.0 | 9375 | 0.1057 | 0.8330 | 0.8738 | 0.9712 | 0.8529 |
|
| 81 |
-
| 0.0325 | 6.0 | 11250 | 0.1103 | 0.8585 | 0.8743 | 0.9738 | 0.8663 |
|
| 82 |
-
| 0.0253 | 7.0 | 13125 | 0.1204 | 0.8453 | 0.8825 | 0.9735 | 0.8635 |
|
| 83 |
-
| 0.0203 | 8.0 | 15000 | 0.1247 | 0.8547 | 0.8825 | 0.9741 | 0.8684 |
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
### Framework versions
|
| 87 |
-
|
| 88 |
-
- Transformers 4.19.2
|
| 89 |
-
- Pytorch 1.11.0+cu113
|
| 90 |
-
- Datasets 2.2.2
|
| 91 |
-
- Tokenizers 0.12.1
|
|
|
|
| 32 |
- Accuracy: 0.9741
|
| 33 |
- F1: 0.8684
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
### Use a pipeline as a high-level helper
|
| 36 |
```python
|
| 37 |
from transformers import pipeline
|
|
|
|
| 46 |
tokenizer = AutoTokenizer.from_pretrained("Azma-AI/bert-uncased-keyword-extractor")
|
| 47 |
model = AutoModelForTokenClassification.from_pretrained("Azma-AI/bert-uncased-keyword-extractor")
|
| 48 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|