---
library_name: transformers
license: apache-2.0
base_model: bert-base-cased
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: bert-base-cased-finetuned-AddedTokens-HMGCR-IC50s-V1
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# bert-base-cased-finetuned-AddedTokens-HMGCR-IC50s-V1

This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on 905 HMGCR 
IC50 values from bindingDB.org. Molecules with counter ions were included twice, once with and once without counter-ions.

It achieves the following results on the evaluation set:
- Loss: 0.7278
- Accuracy: 0.7929
- F1: 0.7931

## Model description

More information needed

## Intended uses & limitations

Can classify HMGCR IC50 values as < 50 nM, < 500 nM, and > 500 nM. See Confusion matrix below:


![image/png](https://cdn-uploads.huggingface.co/production/uploads/679e079d375d81eb7ca4850e/VlV7AHGYSNKi2n5l62oGl.png)

## 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: 32
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 20

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1     |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.9314        | 1.0   | 25   | 0.8466          | 0.7071   | 0.6371 |
| 0.7535        | 2.0   | 50   | 0.7025          | 0.7357   | 0.6634 |
| 0.6292        | 3.0   | 75   | 0.6237          | 0.7714   | 0.6956 |
| 0.5464        | 4.0   | 100  | 0.6162          | 0.7571   | 0.7137 |
| 0.5068        | 5.0   | 125  | 0.5730          | 0.7857   | 0.7185 |
| 0.4516        | 6.0   | 150  | 0.5872          | 0.7643   | 0.7312 |
| 0.3971        | 7.0   | 175  | 0.6004          | 0.7643   | 0.7578 |
| 0.3768        | 8.0   | 200  | 0.6253          | 0.7714   | 0.7739 |
| 0.3353        | 9.0   | 225  | 0.6280          | 0.7786   | 0.7522 |
| 0.3439        | 10.0  | 250  | 0.6299          | 0.7714   | 0.7613 |
| 0.3087        | 11.0  | 275  | 0.6569          | 0.7786   | 0.7719 |
| 0.2979        | 12.0  | 300  | 0.6308          | 0.7714   | 0.7753 |
| 0.2561        | 13.0  | 325  | 0.6596          | 0.7786   | 0.7786 |
| 0.2703        | 14.0  | 350  | 0.6646          | 0.7786   | 0.7808 |
| 0.2504        | 15.0  | 375  | 0.7125          | 0.7857   | 0.7913 |
| 0.2397        | 16.0  | 400  | 0.6893          | 0.7786   | 0.7770 |
| 0.2152        | 17.0  | 425  | 0.7278          | 0.7929   | 0.7931 |
| 0.2066        | 18.0  | 450  | 0.6947          | 0.7857   | 0.7895 |
| 0.2133        | 19.0  | 475  | 0.7202          | 0.7714   | 0.7756 |
| 0.202         | 20.0  | 500  | 0.7167          | 0.7857   | 0.7887 |


### Framework versions

- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.3.0
- Tokenizers 0.21.0