README.md

metadata

tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: libCap_prBERTbfd_clf
    results: []

libCap_prBERTbfd_clf

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

• Loss: 0.5197
• Accuracy: 0.7457
• F1: 0.7940
• Precision: 0.7567
• Recall: 0.8352
• Auroc: 0.7268

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-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 4096
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	Auroc
No log	0.98	34	0.6393	0.6396	0.7053	0.6782	0.7345	0.6197
No log	1.98	68	0.5713	0.6962	0.7499	0.7256	0.7759	0.6795
No log	2.98	102	0.5652	0.7126	0.7388	0.7918	0.6924	0.7168
No log	3.98	136	0.5360	0.7330	0.7896	0.7345	0.8536	0.7076
No log	4.98	170	0.5223	0.7423	0.7830	0.7740	0.7921	0.7318
No log	5.98	204	0.5180	0.7454	0.7882	0.7699	0.8075	0.7323
No log	6.98	238	0.5179	0.7440	0.7934	0.7537	0.8376	0.7243
No log	7.98	272	0.5197	0.7457	0.7940	0.7567	0.8352	0.7268

Framework versions

• Transformers 4.21.1
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1