README.md

metadata

base_model: KB/bert-base-swedish-cased
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: news_category_classification
    results: []

news_category_classification

This model is a fine-tuned version of KB/bert-base-swedish-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.8615
• Accuracy: 0.7286
• F1: 0.7300
• Precision: 0.7351
• Recall: 0.7286
• Accuracy Label Arts, culture, entertainment and media: 0.8333
• Accuracy Label Conflict, war and peace: 0.7234
• Accuracy Label Crime, law and justice: 0.7919
• Accuracy Label Disaster, accident, and emergency incident: 0.8931
• Accuracy Label Economy, business, and finance: 0.7975
• Accuracy Label Environment: 0.4375
• Accuracy Label Health: 0.7
• Accuracy Label Human interest: 0.3333
• Accuracy Label Labour: 0.5
• Accuracy Label Lifestyle and leisure: 0.5
• Accuracy Label Politics: 0.6331
• Accuracy Label Religion: 0.0
• Accuracy Label Science and technology: 0.4167
• Accuracy Label Society: 0.4561
• Accuracy Label Sport: 0.9615
• Accuracy Label Weather: 1.0

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	Accuracy Label Arts, culture, entertainment and media	Accuracy Label Conflict, war and peace	Accuracy Label Crime, law and justice	Accuracy Label Disaster, accident, and emergency incident	Accuracy Label Economy, business, and finance	Accuracy Label Environment	Accuracy Label Health	Accuracy Label Human interest	Accuracy Label Labour	Accuracy Label Lifestyle and leisure	Accuracy Label Politics	Accuracy Label Religion	Accuracy Label Science and technology	Accuracy Label Society	Accuracy Label Sport	Accuracy Label Weather
1.7671	0.3373	200	1.5661	0.5554	0.5206	0.5828	0.5554	0.5833	0.7553	0.8960	0.3206	0.6709	0.125	0.7	0.0	0.5	0.5	0.2878	0.0	0.0	0.0351	0.9615	1.0
1.0248	0.6745	400	1.0774	0.6709	0.6591	0.6984	0.6709	0.9167	0.7979	0.8150	0.8626	0.7215	0.375	0.9	0.25	1.0	0.5	0.3094	0.0	0.4167	0.1930	0.9615	1.0
0.5845	1.0118	600	0.9907	0.6536	0.6563	0.6829	0.6536	0.9167	0.7287	0.6763	0.8779	0.7215	0.4375	0.8	0.0	1.0	0.75	0.3669	0.0	0.4167	0.4386	0.9231	1.0
0.6104	1.3491	800	0.8674	0.7240	0.7233	0.7333	0.7240	0.8333	0.7021	0.8324	0.8779	0.7848	0.5	0.7	0.25	1.0	0.75	0.6331	0.0	0.25	0.3684	0.9615	1.0
0.4223	1.6863	1000	0.8602	0.7240	0.7250	0.7387	0.7240	0.75	0.6755	0.8844	0.8550	0.7342	0.5	0.9	0.3333	1.0	0.625	0.6475	0.0	0.3333	0.3684	0.9615	0.0
0.3104	2.0236	1200	0.8565	0.7263	0.7266	0.7326	0.7263	0.8333	0.7181	0.8324	0.9084	0.7722	0.4375	0.7	0.25	0.5	0.75	0.5612	0.0	0.4167	0.4737	0.9615	1.0
0.2855	2.3609	1400	0.8981	0.7240	0.7283	0.7402	0.7240	0.75	0.7394	0.8324	0.8550	0.7975	0.5	0.7	0.3333	0.5	0.625	0.5899	0.0	0.4167	0.3860	0.9615	1.0
0.217	2.6981	1600	0.8667	0.7309	0.7292	0.7358	0.7309	0.75	0.7447	0.8382	0.8931	0.8481	0.375	0.8	0.3333	0.5	0.5	0.5396	0.0	0.4167	0.4561	0.9615	1.0

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1