premsa/political-bias-prediction-allsides-mBERT

license: apache-2.0 ---

model base: https://huggingface.co/google-bert/bert-base-multilingual-uncased

dataset: https://github.com/ramybaly/Article-Bias-Prediction

training parameters:

• —batch_size: 100
• —epochs: 5
• —dropout: 0.05
• —max_length: 512
• —learning_rate: 3e-5
• —warmup_steps: 100
• —random_state: 239

training methodology:

• —sanitize dataset following specific rule-set, utilize random split as provided in the dataset
• —train on train split and evaluate on validation split in each epoch
• —evaluate test split only on the model that performed best on validation loss

result summary:

• —throughout the five training epochs, model of second epoch achieved the lowest validation loss of 0.3003
• —on test split second epoch model achieved f1 score of 0.8842

usage:

model = AutoModelForSequenceClassification.from_pretrained("premsa/political-bias-prediction-allsides-mBERT")
tokenizer = AutoTokenizer.from_pretrained(premsa/"political-bias-prediction-allsides-mBERT")
nlp = pipeline("text-classification", model=model, tokenizer=tokenizer)
print(nlp("die massen werden von den medien kontrolliert."))