metadata

widget:
  - text: The reduction of carbon emissions is improving for the last 2 years.
    example_title: Example 1
    candidate_labels: Related to Environmental Claims, Not related to Environmental Claims
  - text: The weather is very sunny today.
    example_title: Example 2
language: en
datasets:
  - climatebert/environmental_claims
tags:
  - Text Classification
  - environmental-claims
  - bert-base-uncased
model-index:
  - name: Vinoth24/environmental_claims
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: environmental-claims
          type: environmental-claims
          config: environmental-claims
          split: validation & test
        metrics:
          - name: Loss
            type: loss
            value: 0.4887

Model Card for environmental-claims

Model Description

The environmental-claims model is fine-tuned using the EnvironmentalClaims dataset on Bert base-uncased model. This model is fine-tuned with the help of Happy Transformers on the Bert base-uncased model. The EnvironmentalClaims dataset is annotated by finance and sustainable finance students and authors of Zurich University. This model is expected to predict whether the input sequence is related to real-time environmental claims or not.

Usage

loading the model :

from happytransformer import HappyTextClassification
happy_class = HappyTextClassification(model_type="BERT", model_name="Vinoth24/environmental_claims")

prediction :

result = happy_class.classify_text('The reduction of carbon emissions is improving for the last 2 years.')
print(result) -- TextClassificationResult(label='LABEL_1', score=0.9948860359191895)
print(result.label) -- LABEL_1
print(result.score) -- 0.994

Result Interpretation:

LABEL_1 - Related to Environmental Claims
LABEL_0 - Not Related to Environmental Claims

Feel free to train the model more with your custom Environmental claims data. Any queries will be answered.
Thank you! :)

Created by Kasi Vinoth S from India