bert-base-german-cased-amdi-synset

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5931
• Accuracy: 0.8296
• F1: 0.6552
• Precision: 0.6671
• Recall: 0.6838

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: 32
• 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: 50
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
3.8594	0.3049	50	3.1098	0.3546	0.1123	0.1516	0.1536
2.0232	0.6098	100	1.3093	0.6799	0.3477	0.3439	0.4111
1.2061	0.9146	150	1.0550	0.7005	0.4147	0.4034	0.4774
0.8326	1.2195	200	0.8302	0.7504	0.4769	0.5008	0.5181
0.7385	1.5244	250	0.7518	0.7659	0.5069	0.5359	0.5543
0.6504	1.8293	300	0.7083	0.7762	0.5155	0.4996	0.5648
0.6269	2.1341	350	0.6032	0.8176	0.5909	0.5914	0.6244
0.4735	2.4390	400	0.6070	0.8090	0.6165	0.6480	0.6377
0.4269	2.7439	450	0.6315	0.8090	0.6380	0.6571	0.6666
0.4783	3.0488	500	0.5931	0.8296	0.6552	0.6671	0.6838
0.3407	3.3537	550	0.5612	0.8382	0.6595	0.6541	0.6934
0.3022	3.6585	600	0.5809	0.8262	0.6694	0.6828	0.6933
0.3161	3.9634	650	0.5659	0.8434	0.6834	0.6953	0.7053
0.2405	4.2683	700	0.6109	0.8382	0.6643	0.6651	0.6965
0.2201	4.5732	750	0.5762	0.8485	0.6880	0.6913	0.7115
0.2188	4.8780	800	0.5860	0.8485	0.6875	0.6911	0.7129
0.16	5.1829	850	0.6092	0.8399	0.6630	0.6681	0.6882
0.1456	5.4878	900	0.6303	0.8417	0.6646	0.6718	0.6873
0.1718	5.7927	950	0.6210	0.8468	0.6703	0.6734	0.6947

Framework versions

• Transformers 4.45.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.20.3