furina_seed42_eng_kin_hau

This model is a fine-tuned version of yihongLiu/furina on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0202
• Spearman Corr: 0.7925

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	0.59	200	0.0427	0.4859
No log	1.18	400	0.0345	0.6546
No log	1.77	600	0.0295	0.6863
0.0527	2.36	800	0.0272	0.6929
0.0527	2.95	1000	0.0284	0.7195
0.0527	3.55	1200	0.0271	0.7192
0.0244	4.14	1400	0.0253	0.7366
0.0244	4.73	1600	0.0236	0.7434
0.0244	5.32	1800	0.0239	0.7436
0.0244	5.91	2000	0.0255	0.7483
0.017	6.5	2200	0.0230	0.7575
0.017	7.09	2400	0.0231	0.7541
0.017	7.68	2600	0.0230	0.7568
0.0123	8.27	2800	0.0228	0.7575
0.0123	8.86	3000	0.0233	0.7633
0.0123	9.45	3200	0.0228	0.7679
0.0092	10.04	3400	0.0226	0.7647
0.0092	10.64	3600	0.0220	0.7704
0.0092	11.23	3800	0.0214	0.7717
0.0092	11.82	4000	0.0219	0.7768
0.0074	12.41	4200	0.0215	0.7760
0.0074	13.0	4400	0.0209	0.7792
0.0074	13.59	4600	0.0206	0.7796
0.006	14.18	4800	0.0211	0.7770
0.006	14.77	5000	0.0211	0.7801
0.006	15.36	5200	0.0216	0.7807
0.006	15.95	5400	0.0205	0.7841
0.0052	16.54	5600	0.0211	0.7846
0.0052	17.13	5800	0.0206	0.7873
0.0052	17.73	6000	0.0202	0.7863
0.0045	18.32	6200	0.0205	0.7858
0.0045	18.91	6400	0.0202	0.7886
0.0045	19.5	6600	0.0208	0.7876
0.0042	20.09	6800	0.0203	0.7902
0.0042	20.68	7000	0.0211	0.7850
0.0042	21.27	7200	0.0204	0.7899
0.0042	21.86	7400	0.0207	0.7905
0.0037	22.45	7600	0.0202	0.7925

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1