README.md

metadata

datasets:
  - Mnwa/russian-toxic
language:
  - ru
metrics:
  - precision
  - recall
  - f1
model-index:
  - name: toxic-detector-roberta
    results:
      - task:
          name: Toxic-detector-roberta
          type: text-classification
        dataset:
          name: Mnwa/russian-toxic
          type: Mnwa/russian-toxic
          split: test
          args:
            language: ru
        metrics:
          - name: precision class negative
            type: precision
            value: 0.92
          - name: precision class positive
            type: precision
            value: 0.97
          - name: recall class negative
            type: recall
            value: 0.9
          - name: recall class positive
            type: recall
            value: 0.98
          - name: f1 class negative
            type: f1
            value: 0.91
          - name: f1 class positive
            type: f1
            value: 0.98
pipeline_tag: text-classification

model pretrained in dataset Mnwa/russian-toxic after finetuned