furina_seed42_eng_kin_amh_basic_0.0001

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

• Loss: 0.0286
• Spearman Corr: 0.7423

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: 0.0001
• 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	1.75	200	0.0352	0.5952
0.0595	3.51	400	0.0331	0.6921
0.0223	5.26	600	0.0303	0.7036
0.013	7.02	800	0.0320	0.7059
0.0094	8.77	1000	0.0320	0.7261
0.0067	10.53	1200	0.0331	0.7353
0.005	12.28	1400	0.0293	0.7446
0.0042	14.04	1600	0.0251	0.7440
0.0042	15.79	1800	0.0284	0.7340
0.0036	17.54	2000	0.0306	0.7380
0.0033	19.3	2200	0.0267	0.7394
0.0029	21.05	2400	0.0269	0.7404
0.0026	22.81	2600	0.0279	0.7402
0.0024	24.56	2800	0.0278	0.7418
0.0022	26.32	3000	0.0288	0.7384
0.002	28.07	3200	0.0286	0.7423

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2