a919d12fc4247472475dde5e2c1ae210

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 1.3250
• Data Size: 1.0
• Epoch Runtime: 88.5870
• Accuracy: 0.9214
• F1 Macro: 0.8700

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	37.9805	0	4.5746	0.0877	0.0409
No log	1	500	20.9561	0.0078	5.1706	0.2752	0.1509
No log	2	1000	6.2515	0.0156	6.1480	0.4536	0.1919
No log	3	1500	3.9762	0.0312	8.0494	0.7198	0.5735
No log	4	2000	2.1461	0.0625	11.5318	0.8634	0.8216
0.2596	5	2500	2.5726	0.125	17.1712	0.8569	0.7998
1.1401	6	3000	0.9744	0.25	27.1979	0.9128	0.8694
0.1519	7	3500	1.0809	0.5	47.7444	0.9183	0.8772
0.6546	8.0	4000	0.6451	1.0	89.8200	0.9320	0.8771
0.6903	9.0	4500	0.7754	1.0	88.7658	0.9189	0.8741
0.4732	10.0	5000	0.8694	1.0	89.0307	0.9299	0.8931
0.3565	11.0	5500	1.2469	1.0	87.8701	0.9209	0.8824
0.5164	12.0	6000	1.3250	1.0	88.5870	0.9214	0.8700

Framework versions

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