bert-base-german-cased-noisy-pretrain-fine-tuned_v1.1

This model is a fine-tuned version of tbosse/bert-base-german-cased-finetuned-subj_preTrained_with_noisyData_v1.1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2742
• Precision: 0.8072
• Recall: 0.7657
• F1: 0.7859
• Accuracy: 0.9217

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	0.3112	0.7601	0.5886	0.6634	0.8773
No log	2.0	66	0.2539	0.7706	0.72	0.7445	0.9038
No log	3.0	99	0.2416	0.7755	0.76	0.7677	0.9130
No log	4.0	132	0.2536	0.8190	0.7371	0.7759	0.9165
No log	5.0	165	0.2644	0.7982	0.7457	0.7710	0.9176
No log	6.0	198	0.2735	0.8142	0.7514	0.7816	0.9205
No log	7.0	231	0.2742	0.8072	0.7657	0.7859	0.9217

Framework versions

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