d687edbb826372181c176b80b68505da

This model is a fine-tuned version of distilbert/distilbert-base-german-cased on the nyu-mll/glue [stsb] dataset. It achieves the following results on the evaluation set:

• Loss: 0.7909
• Data Size: 1.0
• Epoch Runtime: 6.2167
• Mse: 0.7911
• Mae: 0.6793
• R2: 0.6461

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	Mse	Mae	R2
No log	0	0	7.1487	0	0.9818	7.1499	2.2408	-2.1984
No log	1	179	4.8580	0.0078	1.2031	4.8591	1.8216	-1.1737
No log	2	358	2.5491	0.0156	1.2582	2.5500	1.3722	-0.1407
No log	3	537	2.3839	0.0312	1.4709	2.3846	1.2922	-0.0667
No log	4	716	1.9508	0.0625	1.6132	1.9514	1.1653	0.1271
No log	5	895	1.1242	0.125	1.9246	1.1243	0.8326	0.4971
0.1171	6	1074	0.8787	0.25	2.8866	0.8790	0.7352	0.6068
0.9116	7	1253	0.8112	0.5	3.9745	0.8115	0.7147	0.6370
0.7932	8.0	1432	0.7665	1.0	6.3199	0.7666	0.6729	0.6571
0.5257	9.0	1611	0.7647	1.0	6.2143	0.7649	0.6832	0.6578
0.3798	10.0	1790	0.7595	1.0	6.1731	0.7597	0.6914	0.6601
0.2764	11.0	1969	0.7619	1.0	6.1021	0.7621	0.6743	0.6591
0.2142	12.0	2148	0.7726	1.0	6.2589	0.7729	0.6859	0.6543
0.1823	13.0	2327	0.8017	1.0	6.1330	0.8018	0.6707	0.6413
0.1622	14.0	2506	0.7909	1.0	6.2167	0.7911	0.6793	0.6461

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1