dc6479deddad0b0cd5436394dc62e44d

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

• Loss: 1.5604
• Data Size: 1.0
• Epoch Runtime: 42.6801
• 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	1.8362	0	2.2665	0.1119	0.0335
No log	1	500	1.7068	0.0078	2.7066	0.2908	0.0751
No log	2	1000	1.6554	0.0156	3.2557	0.3488	0.0862
No log	3	1500	1.6196	0.0312	4.1491	0.3488	0.0862
No log	4	2000	1.6054	0.0625	5.6282	0.3488	0.0862
0.0874	5	2500	1.5693	0.125	8.2687	0.2908	0.0751
1.5991	6	3000	1.5706	0.25	13.3498	0.2908	0.0751
0.2598	7	3500	1.5747	0.5	23.5435	0.3488	0.0862
1.5884	8.0	4000	1.5610	1.0	45.2155	0.3488	0.0862
1.5911	9.0	4500	1.5613	1.0	47.0723	0.3488	0.0862
1.5992	10.0	5000	1.5647	1.0	44.9919	0.3488	0.0862
1.5846	11.0	5500	1.5589	1.0	45.5251	0.3488	0.0862
1.5618	12.0	6000	1.5644	1.0	45.6571	0.3488	0.0862
1.5664	13.0	6500	1.5564	1.0	44.0369	0.3488	0.0862
1.5871	14.0	7000	1.5620	1.0	46.1823	0.3488	0.0862
1.5816	15.0	7500	1.5676	1.0	45.9794	0.2908	0.0751
1.5841	16.0	8000	1.5614	1.0	45.0035	0.3488	0.0862
1.5662	17.0	8500	1.5604	1.0	42.6801	0.3488	0.0862

Framework versions

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