kitrofimov/news-clf

Classifies English-language news into 4 categories: world, sports, business, and science / technology.

Fine-tuned version of distilbert/distilbert-base-uncased on fancyzhx/ag_news, trained for 3 epochs with 128 token truncation.

It achieves the following results on the evaluation set:

• Loss: 0.1759
• Accuracy: 0.9414

Made as a homework project for 4th lesson of the FastAI's Practical Deep Learning for Coders course. Hugging Face Spaces demo available here.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.2009	1.0	938	0.1998	0.9307
0.1773	2.0	1876	0.1804	0.9375
0.1418	3.0	2814	0.1759	0.9414

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.1.0
• Tokenizers 0.22.0