bert-base-german-cased-noisy-pretrain-fine-tuned_v2

This model is a fine-tuned version of tbosse/bert-base-german-cased-finetuned-subj_preTrained_with_noisyData_v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2872
• Precision: 0.7870
• Recall: 0.76
• F1: 0.7733
• Accuracy: 0.9159

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: 7

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	33	0.3105	0.7731	0.5743	0.6590	0.8813
No log	2.0	66	0.2632	0.7588	0.7371	0.7478	0.9055
No log	3.0	99	0.2517	0.7630	0.7543	0.7586	0.9096
No log	4.0	132	0.2590	0.8145	0.74	0.7754	0.9171
No log	5.0	165	0.2665	0.7939	0.7486	0.7706	0.9165
No log	6.0	198	0.2854	0.7951	0.7429	0.7681	0.9147
No log	7.0	231	0.2872	0.7870	0.76	0.7733	0.9159

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1