640c520a803255939117ca9f7e1b12b8

This model is a fine-tuned version of google-bert/bert-base-german-cased on the contemmcm/amazon_reviews_2013 [cell-phone] dataset. It achieves the following results on the evaluation set:

• Loss: 0.9511
• Data Size: 1.0
• Epoch Runtime: 109.5808
• Accuracy: 0.6533
• F1 Macro: 0.5706

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	1.6333	0	7.2150	0.1092	0.0547
No log	1	1973	1.5115	0.0078	8.4910	0.2809	0.1343
0.0326	2	3946	1.4708	0.0156	9.1047	0.3853	0.1112
1.4199	3	5919	1.3540	0.0312	10.8742	0.4639	0.2224
1.2564	4	7892	1.1363	0.0625	14.4120	0.5342	0.3517
1.0721	5	9865	1.0368	0.125	20.1523	0.5815	0.4252
0.965	6	11838	0.9522	0.25	33.3651	0.6084	0.4646
0.9626	7	13811	0.8878	0.5	57.7552	0.6314	0.5446
0.8541	8.0	15784	0.8571	1.0	108.5909	0.6412	0.5747
0.7655	9.0	17757	0.8673	1.0	108.2613	0.6509	0.5560
0.6928	10.0	19730	0.8930	1.0	107.9227	0.6340	0.5778
0.6262	11.0	21703	0.9613	1.0	108.4263	0.6426	0.5554
0.5832	12.0	23676	0.9511	1.0	109.5808	0.6533	0.5706

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1