Text Classification
Transformers
Safetensors
Russian
bert
icd10
medical
knowledge-distillation-teacher
text-embeddings-inference
Instructions to use Dmitry43243242/icd10-ru-chapter-teacher-cointegrated_rubert_tiny2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Dmitry43243242/icd10-ru-chapter-teacher-cointegrated_rubert_tiny2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="Dmitry43243242/icd10-ru-chapter-teacher-cointegrated_rubert_tiny2", device_map="auto")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("Dmitry43243242/icd10-ru-chapter-teacher-cointegrated_rubert_tiny2") model = AutoModelForSequenceClassification.from_pretrained("Dmitry43243242/icd10-ru-chapter-teacher-cointegrated_rubert_tiny2", device_map="auto") - Notebooks
- Google Colab
- Kaggle
ICD-10 Russian Chapter Teacher: cointegrated/rubert-tiny2
Fine-tuned BERT-family teacher model for Russian ICD-10 chapter/group prediction.
This model is exported from train_extended_models.ipynb as teacher #4 of top-5, selected by validation Top-3 accuracy.
- Base model:
cointegrated/rubert-tiny2 - Repo:
Dmitry43243242/icd10-ru-chapter-teacher-cointegrated_rubert_tiny2 - Labels: ICD-10 chapter letters, excluding U
- Intended use: teacher model for knowledge distillation and top-3 decision support
Metrics
| Split | Top-1 | Top-3 | F1 macro | F1 weighted |
|---|---|---|---|---|
| Validation | 0.6719 | 0.8637 | 0.3590 | 0.6638 |
| Test | 0.6866 | 0.8556 | 0.3530 | 0.6797 |
This is a decision-support model, not an automatic diagnosis system.
- Downloads last month
- 2