DaInternet12/bert_affinity_prediction

results

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

• Loss: 0.3504
• Mse: 0.3504
• Mae: 0.7117

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Mse	Mae
0.4084	0.9091	20	0.3711	0.3711	0.7489
0.3852	1.8182	40	0.5426	0.5426	0.9613
0.4357	2.7273	60	0.6162	0.6162	0.9952
0.4939	3.6364	80	0.4569	0.4569	0.8551
0.4794	4.5455	100	0.4929	0.4929	0.9021
0.3671	5.4545	120	0.4535	0.4535	0.8575
0.4015	6.3636	140	0.3906	0.3906	0.7767
0.4432	7.2727	160	0.4684	0.4684	0.8312
0.3985	8.1818	180	0.4389	0.4389	0.8397
0.3911	9.0909	200	0.4122	0.4122	0.8004
0.3856	10.0	220	0.3956	0.3956	0.7809

Framework versions

• Transformers 4.40.2
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.19.1