furina_original_esp-kin-eng_train_spearman_corr

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

• Loss: 0.0243
• Spearman Corr: 0.7225

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	1.63	200	0.0285	0.5491
0.0764	3.27	400	0.0332	0.6502
0.0248	4.9	600	0.0243	0.6852
0.0185	6.53	800	0.0217	0.7121
0.0148	8.16	1000	0.0270	0.7180
0.0148	9.8	1200	0.0243	0.7170
0.0121	11.43	1400	0.0240	0.7203
0.01	13.06	1600	0.0250	0.7205
0.0083	14.69	1800	0.0251	0.7242
0.0073	16.33	2000	0.0243	0.7153
0.0073	17.96	2200	0.0257	0.7238
0.0064	19.59	2400	0.0235	0.7228
0.0058	21.22	2600	0.0246	0.7193
0.0054	22.86	2800	0.0246	0.7221
0.0051	24.49	3000	0.0235	0.7222
0.0048	26.12	3200	0.0236	0.7224
0.0048	27.76	3400	0.0243	0.7225

Framework versions

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