furina_afr_corr_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: 0.0213
• Spearman Corr: 0.7660

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	0.85	200	0.0813	0.0483
No log	1.69	400	0.0439	0.3738
0.1125	2.54	600	0.0360	0.6162
0.1125	3.38	800	0.0287	0.6660
0.035	4.23	1000	0.0271	0.6987
0.035	5.07	1200	0.0248	0.7139
0.035	5.92	1400	0.0252	0.7230
0.026	6.77	1600	0.0246	0.7173
0.026	7.61	1800	0.0230	0.7344
0.0227	8.46	2000	0.0236	0.7361
0.0227	9.3	2200	0.0223	0.7470
0.0203	10.15	2400	0.0227	0.7462
0.0203	10.99	2600	0.0224	0.7477
0.0203	11.84	2800	0.0242	0.7483
0.0186	12.68	3000	0.0218	0.7537
0.0186	13.53	3200	0.0229	0.7528
0.0175	14.38	3400	0.0227	0.7552
0.0175	15.22	3600	0.0221	0.7602
0.0163	16.07	3800	0.0216	0.7612
0.0163	16.91	4000	0.0216	0.7619
0.0163	17.76	4200	0.0217	0.7616
0.0154	18.6	4400	0.0215	0.7612
0.0154	19.45	4600	0.0212	0.7651
0.0146	20.3	4800	0.0216	0.7634
0.0146	21.14	5000	0.0213	0.7641
0.0141	21.99	5200	0.0223	0.7629
0.0141	22.83	5400	0.0213	0.7664
0.0141	23.68	5600	0.0214	0.7635
0.0134	24.52	5800	0.0214	0.7649
0.0134	25.37	6000	0.0216	0.7637
0.013	26.22	6200	0.0214	0.7644
0.013	27.06	6400	0.0213	0.7656
0.013	27.91	6600	0.0214	0.7661
0.0127	28.75	6800	0.0212	0.7655
0.0127	29.6	7000	0.0213	0.7660

Framework versions

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