README.md

metadata

license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - emotion
metrics:
  - accuracy
model-index:
  - name: twitter-emotion-deberta-v3-base
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: emotion
          type: emotion
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.937
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: emotion
          type: emotion
          config: default
          split: test
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9255
            verified: true
          - name: Precision Macro
            type: precision
            value: 0.8915483806374028
            verified: true
          - name: Precision Micro
            type: precision
            value: 0.9255
            verified: true
          - name: Precision Weighted
            type: precision
            value: 0.9286522707274408
            verified: true
          - name: Recall Macro
            type: recall
            value: 0.875946770128528
            verified: true
          - name: Recall Micro
            type: recall
            value: 0.9255
            verified: true
          - name: Recall Weighted
            type: recall
            value: 0.9255
            verified: true
          - name: F1 Macro
            type: f1
            value: 0.8790048313120858
            verified: true
          - name: F1 Micro
            type: f1
            value: 0.9255
            verified: true
          - name: F1 Weighted
            type: f1
            value: 0.92449885920049
            verified: true
          - name: loss
            type: loss
            value: 0.16804923117160797
            verified: true

twitter-emotion-deberta-v3-base

This model is a fine-tuned version of DeBERTa-v3. It achieves the following results on the evaluation set:

• Loss: 0.1474
• Accuracy: 0.937

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 80
• eval_batch_size: 80
• lr_scheduler_type: linear
• num_epochs: 6.0

Framework versions

• Transformers 4.12.5
• Pytorch 1.10.0+cu113
• Datasets 1.15.1
• Tokenizers 0.10.3