| --- |
| license: apache-2.0 |
| language: |
| - ru |
| base_model: |
| - google-bert/bert-base-multilingual-cased |
| metrics: |
| - eval_loss |
| - eval_precision |
| - eval_recall |
| - eval_f1 |
| - eval_runtime |
| - eval_samples_per_second |
| - eval_steps_per_second |
|
|
| model-index: |
| - name: bert_realestate_rus |
| results: |
| - task: |
| type: token-classification |
| dataset: |
| type: custom |
| name: own_main_test_dataset |
| metrics: |
| - name: eval_loss |
| type: eval_loss |
| value: 0.04553 |
| verified: false |
| - name: eval_precision |
| type: eval_precision |
| value: 0.99006 |
| verified: false |
| - name: eval_recall |
| type: eval_recall |
| value: 0.99416 |
| verified: false |
| - name: eval_f1 |
| type: eval_f1 |
| value: 0.99211 |
| verified: false |
|
|
| --- |
| |
|
|
| **Model Features:** |
| - Based on BERT architecture. |
| - Fine-tuned and optimized on annotated data in CoNLL format with real estate-specific labels. |
