furina_seed42_eng_kin_amh_roman

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

• Loss: 0.0284
• Spearman Corr: 0.7771

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	0.65	200	0.0373	0.5747
No log	1.3	400	0.0297	0.6851
No log	1.95	600	0.0311	0.7236
0.0545	2.61	800	0.0305	0.7322
0.0545	3.26	1000	0.0281	0.7496
0.0545	3.91	1200	0.0278	0.7582
0.0208	4.56	1400	0.0278	0.7528
0.0208	5.21	1600	0.0238	0.7556
0.0208	5.86	1800	0.0235	0.7631
0.0143	6.51	2000	0.0245	0.7634
0.0143	7.17	2200	0.0243	0.7619
0.0143	7.82	2400	0.0242	0.7651
0.0102	8.47	2600	0.0257	0.7645
0.0102	9.12	2800	0.0271	0.7713
0.0102	9.77	3000	0.0255	0.7661
0.0079	10.42	3200	0.0218	0.7720
0.0079	11.07	3400	0.0250	0.7658
0.0079	11.73	3600	0.0266	0.7628
0.0064	12.38	3800	0.0267	0.7657
0.0064	13.03	4000	0.0261	0.7680
0.0064	13.68	4200	0.0232	0.7720
0.0055	14.33	4400	0.0256	0.7737
0.0055	14.98	4600	0.0237	0.7736
0.0055	15.64	4800	0.0284	0.7771

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.1