b3908e1c32b3a07fe1844b73d55e2a82

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

• Loss: 1.5636
• Data Size: 1.0
• Epoch Runtime: 86.5282
• 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	1.9505	0	3.7013	0.0333	0.0107
No log	1	500	1.6020	0.0078	5.1566	0.3191	0.1253
No log	2	1000	1.7457	0.0156	5.5999	0.3488	0.0862
No log	3	1500	1.6089	0.0312	7.7884	0.3488	0.0862
No log	4	2000	1.6018	0.0625	11.3238	0.2908	0.0751
0.0871	5	2500	1.5823	0.125	17.0985	0.2908	0.0751
1.6062	6	3000	1.5775	0.25	27.9764	0.2908	0.0751
0.2604	7	3500	1.5779	0.5	48.5639	0.3488	0.0862
1.5998	8.0	4000	1.5646	1.0	87.8921	0.2908	0.0751
1.597	9.0	4500	1.5622	1.0	86.6630	0.3488	0.0862
1.6023	10.0	5000	1.5619	1.0	87.7605	0.3488	0.0862
1.5951	11.0	5500	1.5717	1.0	87.2117	0.3488	0.0862
1.5668	12.0	6000	1.5642	1.0	86.8940	0.3488	0.0862
1.5683	13.0	6500	1.5613	1.0	86.3353	0.3488	0.0862
1.5911	14.0	7000	1.5633	1.0	86.0682	0.3488	0.0862
1.5803	15.0	7500	1.5635	1.0	86.9644	0.2908	0.0751
1.5863	16.0	8000	1.5643	1.0	87.6006	0.3488	0.0862
1.5798	17.0	8500	1.5636	1.0	86.5282	0.3488	0.0862

Framework versions

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