File size: 3,305 Bytes
3798ae6
 
 
 
 
59bfe82
 
 
3798ae6
59bfe82
982051b
3798ae6
982051b
3798ae6
982051b
 
 
 
 
3798ae6
 
 
 
 
982051b
3798ae6
46492d4
3798ae6
59bfe82
 
 
 
 
 
 
 
 
 
 
 
3798ae6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
982051b
3798ae6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
982051b
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
84
85
---
library_name: transformers
license: mit
base_model: microsoft/deberta-v3-large
tags:
- multi-label
- question-answering
- text-classification
- generated_from_trainer
datasets:
- saiteki-kai/BeaverTails-it
metrics:
- f1
- accuracy
- precision
- recall
language:
- it
- en
---

<!-- 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. -->

# QA-DeBERTa-v3-large

This model is a fine-tuned version of [microsoft/deberta-v3-large](https://huggingface.co/microsoft/deberta-v3-large) on the [saiteki-kai/BeaverTails-it](https://huggingface.co/datasets/saiteki-kai/BeaverTails-it) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0808
- Accuracy: 0.6938
- Macro F1: 0.6484
- Macro Precision: 0.7149
- Macro Recall: 0.6176
- Micro F1: 0.7545
- Micro Precision: 0.7874
- Micro Recall: 0.7242
- Flagged/accuracy: 0.8566
- Flagged/precision: 0.8975
- Flagged/recall: 0.8380
- Flagged/f1: 0.8667

## 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: 3.85e-06
- train_batch_size: 16
- eval_batch_size: 64
- seed: 42
- distributed_type: multi-GPU
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10

### Training results

| Training Loss | Epoch | Step   | Validation Loss | Accuracy | Macro F1 | Macro Precision | Macro Recall | Micro F1 | Micro Precision | Micro Recall | Flagged/accuracy | Flagged/precision | Flagged/recall | Flagged/f1 |
|:-------------:|:-----:|:------:|:---------------:|:--------:|:--------:|:---------------:|:------------:|:--------:|:---------------:|:------------:|:----------------:|:-----------------:|:--------------:|:----------:|
| 0.0985        | 1.0   | 33814  | 0.0877          | 0.6750   | 0.6102   | 0.6629          | 0.5948       | 0.7406   | 0.7705          | 0.7129       | 0.8447           | 0.8701            | 0.8475         | 0.8586     |
| 0.0867        | 2.0   | 67628  | 0.0817          | 0.6910   | 0.6185   | 0.7559          | 0.5598       | 0.7446   | 0.8165          | 0.6842       | 0.8465           | 0.9093            | 0.8043         | 0.8536     |
| 0.0561        | 3.0   | 101442 | 0.0808          | 0.6938   | 0.6484   | 0.7149          | 0.6177       | 0.7545   | 0.7875          | 0.7242       | 0.8566           | 0.8975            | 0.8380         | 0.8667     |
| 0.0913        | 4.0   | 135256 | 0.0812          | 0.6877   | 0.6412   | 0.7136          | 0.6144       | 0.7516   | 0.7796          | 0.7255       | 0.8546           | 0.8902            | 0.8428         | 0.8658     |
| 0.0709        | 5.0   | 169070 | 0.0826          | 0.6911   | 0.6376   | 0.7306          | 0.5982       | 0.7500   | 0.7911          | 0.7129       | 0.8538           | 0.8936            | 0.8370         | 0.8643     |


### Framework versions

- Transformers 4.51.3
- Pytorch 2.7.0+cu118
- Datasets 3.5.1
- Tokenizers 0.21.1