Bert is a Transformer Bidirectional Encoder based Architecture trained on MLM(Mask Language Modeling) objective. bert-base-uncased finetuned on the poem_sentiment dataset using HuggingFace Trainer with below training parameters.

 learning rate 2e-5, 
 batch size 8,
 num_train_epochs=8,

Model Performance

Epoch	Training Loss	Validation Loss	Accuracy	F1
8	0.468200	0.458632	0.904762	0.899756

How to Use the Model

python

from transformers import pipeline
nlp = pipeline(task='text-classification', 
               model='nickwong64/bert-base-uncased-poems-sentiment')
p1 = "No man is an island, Entire of itself, Every man is a piece of the continent, A part of the main."
p2 = "Ten years, dead and living dim and draw apart. I don’t try to remember, But forgetting is hard."
p3 = "My mind to me a kingdom is; Such present joys therein I find,That it excels all other bliss"
print(nlp(p1))
print(nlp(p2))
print(nlp(p3))
"""
output:
[{'label': 'no_impact', 'score': 0.9982421398162842}]
[{'label': 'negative', 'score': 0.9856176972389221}]
[{'label': 'positive', 'score': 0.9931322932243347}]
"""

Dataset

poem_sentiment

Labels

{0: 'negative', 1: 'positive', 2: 'no_impact', 3: 'mixed'}

Evaluation

{'test_loss': 0.4359096586704254,
 'test_accuracy': 0.9142857142857143,
 'test_f1': 0.9120554830816401,
 'test_runtime': 0.5689,
 'test_samples_per_second': 184.582,
 'test_steps_per_second': 24.611}