furina_ary_corr_2e-05

This model is a fine-tuned version of yihongLiu/furina on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0211
• Spearman Corr: 0.7761

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Spearman Corr
No log	0.9	200	0.0219	0.7722
No log	1.8	400	0.0212	0.7719
0.0017	2.7	600	0.0212	0.7705
0.0017	3.6	800	0.0210	0.7705
0.0016	4.5	1000	0.0212	0.7732
0.0016	5.41	1200	0.0211	0.7715
0.0014	6.31	1400	0.0219	0.7752
0.0014	7.21	1600	0.0213	0.7726
0.0013	8.11	1800	0.0215	0.7717
0.0013	9.01	2000	0.0222	0.7764
0.0013	9.91	2200	0.0212	0.7734
0.0013	10.81	2400	0.0214	0.7714
0.0013	11.71	2600	0.0206	0.7722
0.0012	12.61	2800	0.0222	0.7719
0.0012	13.51	3000	0.0210	0.7723
0.0012	14.41	3200	0.0211	0.7715
0.0012	15.32	3400	0.0206	0.7762
0.0011	16.22	3600	0.0211	0.7761

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2