Marty-17/emotion-classification-model

Model description

This model is a fine-tuned version of distilbert-base-uncased on the dataset dair-ai/emotion dataset. The model is fine tuned to label text into six different emotional categories. These categories being:

• (0) joy, (1) love, (2) anger, (3) fear, (4), surprise (5).

The model achieved the following results below:

• Validation Accuracy: 93.7%
• Test Accuracy: 92%

Intended use

• Analyse text data
• Classify text data as one of the six emotions
• Tested on its ability to classify text into six previously mentioned emotional categories

Training and evaluation data

• The model was trained and evaluated with dair-ai/emotion dataset
• Training set: 16,000 examples
• Validation set: 2,000 examples
• Test set: 2,000 examples

Limitations and Biases

• Model may struggle to understand emotions that require more context.
• Does not capture more nuanced emotions given that are only six emotional categories

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.2517	1.0	1000	0.1897	0.922
0.1476	2.0	2000	0.1656	0.9335

Code Usage Example

  classifier = pipeline("text-classification", model = "Marty-17/emotion-classification-model",tokenizer="Marty-17/emotion-classification-model") # Create Pipeline
    test_text = ["I am so Happy Today !",
                 "I am not Happy Today !",
                 "I do not fear being happy"
                 ] # Test cases

    result = classifier(test_text)
     #Loop and display results for the test case.
    for i,text in enumerate(test_text):
        label= result[i]['label']
        score = result[i]['score']
        print(f"Text Example:{text}")
        print(f"Predicted Label: {label}")