furina_eng_loss_5e-06

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

• Loss: 0.0217
• Spearman Corr: 0.7631

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-06
• 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	1.33	200	0.0222	0.7622
0.0037	2.66	400	0.0226	0.7631
0.0048	3.99	600	0.0222	0.7633
0.0048	5.32	800	0.0227	0.7578
0.0044	6.64	1000	0.0218	0.7644
0.0041	7.97	1200	0.0222	0.7626
0.0041	9.3	1400	0.0219	0.7592
0.004	10.63	1600	0.0217	0.7639
0.0039	11.96	1800	0.0221	0.7634
0.0039	13.29	2000	0.0217	0.7579
0.0038	14.62	2200	0.0221	0.7643
0.0036	15.95	2400	0.0220	0.7617
0.0036	17.28	2600	0.0215	0.7622
0.0035	18.6	2800	0.0220	0.7630
0.0034	19.93	3000	0.0220	0.7645
0.0034	21.26	3200	0.0217	0.7634
0.0034	22.59	3400	0.0216	0.7627
0.0033	23.92	3600	0.0219	0.7642
0.0033	25.25	3800	0.0214	0.7648
0.0032	26.58	4000	0.0216	0.7648
0.0032	27.91	4200	0.0217	0.7631

Framework versions

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