File size: 3,474 Bytes
e09a012
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fdbda3a
 
 
 
 
 
 
e09a012
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fdbda3a
 
e09a012
 
 
fdbda3a
e09a012
 
 
 
fdbda3a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e09a012
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
---
license: cc-by-nc-sa-4.0
base_model: InstaDeepAI/nucleotide-transformer-v2-250m-multi-species
tags:
- generated_from_trainer
metrics:
- precision
- recall
- accuracy
model-index:
- name: nucleotide-transformer-v2-250m-multi-species_ft_BioS73_1kbpHG19_DHSs_H3K27AC
  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. -->

# nucleotide-transformer-v2-250m-multi-species_ft_BioS73_1kbpHG19_DHSs_H3K27AC

This model is a fine-tuned version of [InstaDeepAI/nucleotide-transformer-v2-250m-multi-species](https://huggingface.co/InstaDeepAI/nucleotide-transformer-v2-250m-multi-species) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5967
- F1 Score: 0.8904
- Precision: 0.8843
- Recall: 0.8966
- Accuracy: 0.8822
- Auc: 0.9398
- Prc: 0.9372

## 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: 1e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | F1 Score | Precision | Recall | Accuracy | Auc    | Prc    |
|:-------------:|:------:|:----:|:---------------:|:--------:|:---------:|:------:|:--------:|:------:|:------:|
| 0.4839        | 0.1864 | 500  | 0.4577          | 0.8288   | 0.7331    | 0.9532 | 0.7898   | 0.8989 | 0.8988 |
| 0.3857        | 0.3727 | 1000 | 0.3786          | 0.8651   | 0.8327    | 0.9001 | 0.8502   | 0.9197 | 0.9135 |
| 0.3682        | 0.5591 | 1500 | 0.3529          | 0.8689   | 0.8543    | 0.8841 | 0.8576   | 0.9310 | 0.9303 |
| 0.362         | 0.7454 | 2000 | 0.3569          | 0.8709   | 0.8802    | 0.8617 | 0.8636   | 0.9355 | 0.9330 |
| 0.3336        | 0.9318 | 2500 | 0.4081          | 0.8779   | 0.8023    | 0.9693 | 0.8561   | 0.9397 | 0.9373 |
| 0.2989        | 1.1182 | 3000 | 0.4118          | 0.8799   | 0.8723    | 0.8876 | 0.8707   | 0.9424 | 0.9392 |
| 0.2686        | 1.3045 | 3500 | 0.3579          | 0.8852   | 0.8569    | 0.9155 | 0.8733   | 0.9393 | 0.9343 |
| 0.2696        | 1.4909 | 4000 | 0.3695          | 0.8733   | 0.8957    | 0.8520 | 0.8681   | 0.9439 | 0.9417 |
| 0.2678        | 1.6772 | 4500 | 0.4433          | 0.8944   | 0.8756    | 0.9141 | 0.8848   | 0.9416 | 0.9379 |
| 0.2744        | 1.8636 | 5000 | 0.3976          | 0.8922   | 0.8406    | 0.9504 | 0.8774   | 0.9418 | 0.9394 |
| 0.2166        | 2.0499 | 5500 | 0.5432          | 0.8815   | 0.8802    | 0.8827 | 0.8733   | 0.9426 | 0.9406 |
| 0.1426        | 2.2363 | 6000 | 0.5888          | 0.8988   | 0.8564    | 0.9455 | 0.8863   | 0.9424 | 0.9386 |
| 0.1457        | 2.4227 | 6500 | 0.5963          | 0.8840   | 0.8819    | 0.8862 | 0.8759   | 0.9404 | 0.9370 |
| 0.144         | 2.6090 | 7000 | 0.6159          | 0.8724   | 0.8962    | 0.8499 | 0.8673   | 0.9458 | 0.9454 |
| 0.1279        | 2.7954 | 7500 | 0.5967          | 0.8904   | 0.8843    | 0.8966 | 0.8822   | 0.9398 | 0.9372 |


### Framework versions

- Transformers 4.42.3
- Pytorch 2.3.0+cu121
- Datasets 2.18.0
- Tokenizers 0.19.0