distilbert-base-german-cased-finetuned-tagesschau-subcategories

This model is a fine-tuned version of distilbert-base-german-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5230
• Accuracy: 0.8267

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.4	30	1.5130	0.5733
No log	0.8	60	1.0629	0.7133
No log	1.2	90	0.8431	0.76
No log	1.6	120	0.7812	0.7467
No log	2.0	150	0.6373	0.78
No log	2.4	180	0.5567	0.8133
No log	2.8	210	0.5650	0.8067
No log	3.2	240	0.5068	0.8267
No log	3.6	270	0.5230	0.8267
No log	4.0	300	0.5318	0.8133
No log	4.4	330	0.5327	0.8067
No log	4.8	360	0.4918	0.82

Framework versions

• Transformers 4.25.1
• Pytorch 1.12.1+cu113
• Datasets 2.7.1
• Tokenizers 0.13.2