esm2-t12-35M-lora-64-remote-homology-filtered

This model is a fine-tuned version of facebook/esm2_t12_35M_UR50D on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5657
• Precision: 0.7166
• Recall: 0.6986
• F1-score: 0.7075
• Accuracy: 0.7141

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: 3e-05
• 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_ratio: 0.1
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1-score	Accuracy
0.6191	1.0	7969	0.6185	0.6919	0.5824	0.6325	0.6650
0.5921	2.0	15938	0.5838	0.7201	0.6339	0.6742	0.6968
0.5874	3.0	23907	0.5751	0.7439	0.6104	0.6705	0.7032
0.5593	4.0	31876	0.5664	0.7210	0.6833	0.7016	0.7124
0.576	5.0	39845	0.5657	0.7166	0.6986	0.7075	0.7141

Framework versions

• PEFT 0.11.1
• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2