---
tags:
- generated_from_trainer
datasets:
- setimes
metrics:
- precision
- recall
- f1
- accuracy
language:
- sr
model_index:
- name: distilbert-srb-ner
  results:
  - task:
      name: Token Classification
      type: token-classification
    metric:
      name: Accuracy
      type: accuracy
      value: 0.966779551685212
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# distilbert-srb-ner

This model was finetuned from Aleksandar/distilbert-srb-cased-oscar on the setimes.SR dataset.

It achieves the following results on the evaluation set:
- Loss: 0.1531
- Precision: 0.8296
- Recall: 0.8593
- F1: 0.8442
- Accuracy: 0.9668

## Model description

Finetuned model of Aleksandar/distilbert-srb-cased-oscar on settimes.SR dataset

## Intended uses & limitations

| Tag (IOB)   | Numerical representation | Meaning   (Beginning = B.,  Inside = I.) |
|-------------|--------------------------|------------------------------------------|
| O           | 0                        | Other                                    |
| B-per       | 1                        | B.Person                                 |
| I-per       | 2                        | I. Person                                |
| B-org       | 3                        | B. organization                          |
| I-org       | 4                        | I. organization                          |
| B-loc       | 5                        | B. location                              |
| I-loc       | 6                        | I. location                              |
| B-misc      | 7                        | B. Miscellaneous                         |
| I-misc      | 8                        | I. Miscellaneous                         |
| B-deriv-per | 9                        | B. Derived Person                        |

MIT license

## Training and evaluation data

Training was performed with seqeval as the evaluation metric.  
0.15% of dataset (setimes.SR) was used as validation dataset. 

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10

### Training results

| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| No log        | 1.0   | 207  | 0.2235          | 0.7038    | 0.7314 | 0.7174 | 0.9381   |
| No log        | 2.0   | 414  | 0.1538          | 0.7565    | 0.7937 | 0.7746 | 0.9548   |
| 0.2327        | 3.0   | 621  | 0.1467          | 0.7863    | 0.8171 | 0.8014 | 0.9594   |
| 0.2327        | 4.0   | 828  | 0.1398          | 0.8009    | 0.8526 | 0.8260 | 0.9630   |
| 0.0771        | 5.0   | 1035 | 0.1319          | 0.8021    | 0.8376 | 0.8195 | 0.9637   |
| 0.0771        | 6.0   | 1242 | 0.1482          | 0.8098    | 0.8513 | 0.8300 | 0.9652   |
| 0.0771        | 7.0   | 1449 | 0.1454          | 0.8234    | 0.8510 | 0.8370 | 0.9667   |
| 0.0339        | 8.0   | 1656 | 0.1503          | 0.8310    | 0.8547 | 0.8427 | 0.9663   |
| 0.0339        | 9.0   | 1863 | 0.1525          | 0.8279    | 0.8587 | 0.8430 | 0.9661   |
| 0.0169        | 10.0  | 2070 | 0.1531          | 0.8296    | 0.8593 | 0.8442 | 0.9668   |


### Framework versions

- Transformers 4.9.2
- Pytorch 1.9.0
- Datasets 1.11.0
- Tokenizers 0.10.1