4be5da4a13273b95662f7a02b0747bee

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

• Loss: 0.2213
• Data Size: 1.0
• Epoch Runtime: 48.6894
• Accuracy: 0.9304
• F1 Macro: 0.8766

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	1.8219	0	2.4193	0.0993	0.0601
No log	1	500	1.5722	0.0078	3.1785	0.3589	0.1389
No log	2	1000	1.5065	0.0156	4.0965	0.3503	0.0882
No log	3	1500	1.5120	0.0312	5.9046	0.4612	0.1780
No log	4	2000	1.1079	0.0625	7.0501	0.5842	0.2456
0.0678	5	2500	0.6385	0.125	9.8105	0.7828	0.6156
0.3896	6	3000	0.3752	0.25	15.5161	0.8695	0.8157
0.0379	7	3500	0.2447	0.5	26.9726	0.9138	0.8684
0.1801	8.0	4000	0.1651	1.0	49.3466	0.9289	0.8881
0.1285	9.0	4500	0.1754	1.0	49.2456	0.9254	0.8746
0.1204	10.0	5000	0.1674	1.0	48.8140	0.9340	0.8948
0.0959	11.0	5500	0.1805	1.0	46.6978	0.9259	0.8884
0.0903	12.0	6000	0.2213	1.0	48.6894	0.9304	0.8766

Framework versions

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