c8b130c74f0482ff9eb4b90740c540e6

This model is a fine-tuned version of FacebookAI/xlm-roberta-large-finetuned-conll03-german on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 1.5596
• Data Size: 1.0
• Epoch Runtime: 85.6823
• Accuracy: 0.3488
• F1 Macro: 0.0862

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	1.6597	0	3.6181	0.3337	0.0914
No log	1	500	1.5912	0.0078	4.8214	0.3397	0.1282
No log	2	1000	1.6571	0.0156	5.6805	0.3332	0.0992
No log	3	1500	1.6241	0.0312	7.6939	0.3488	0.0862
No log	4	2000	1.6578	0.0625	11.2933	0.2908	0.0751
0.0873	5	2500	1.5855	0.125	16.8917	0.2908	0.0751
1.6131	6	3000	1.5739	0.25	26.7367	0.2908	0.0751
0.2621	7	3500	1.5719	0.5	47.4582	0.3488	0.0862
1.5933	8.0	4000	1.5668	1.0	85.6839	0.2908	0.0751
1.598	9.0	4500	1.5573	1.0	84.7785	0.3488	0.0862
1.6062	10.0	5000	1.5613	1.0	86.8525	0.3488	0.0862
1.5933	11.0	5500	1.5649	1.0	86.4910	0.3488	0.0862
1.5714	12.0	6000	1.5628	1.0	85.2848	0.3488	0.0862
1.5707	13.0	6500	1.5596	1.0	85.6823	0.3488	0.0862

Framework versions

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