german-party-sentiment-bert-complete-synonyms-5e-5

This model is a fine-tuned version of mdraw/german-news-sentiment-bert on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.8769

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: 5e-05
• train_batch_size: 20
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 120
• num_epochs: 14

Training results

Training Loss	Epoch	Step	Validation Loss
0.9596	1.0	70	0.9676
0.9122	2.0	140	0.8769
0.7382	3.0	210	0.9984
0.5708	4.0	280	1.1080
0.3579	5.0	350	1.4137
0.3066	6.0	420	1.8204
0.1716	7.0	490	1.8167
0.1974	8.0	560	2.1479
0.1164	9.0	630	2.3899
0.0878	10.0	700	2.5266
0.07	11.0	770	2.7014
0.0604	12.0	840	2.7048
0.0278	13.0	910	2.8119
0.0376	14.0	980	2.8799

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Tokenizers 0.15.1