File size: 1,623 Bytes
2679c03
 
 
 
24f8c4f
fd8c65c
2679c03
 
24f8c4f
 
 
 
 
2679c03
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
642b7d0
2679c03
ad87f03
 
8f42a73
 
c028795
ad87f03
2679c03
 
c3eec76
 
 
 
 
 
2679c03
 
 
642b7d0
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
---
tags:
- mt5
- adapter-transformers
- DefinitionGeneration
- definition-modeling
datasets:
- jfeil/GermanDefinitionGeneration-Distillation
license: gpl-3.0
language:
- de
base_model:
- google/mt5-base
---

# Adapter `jfeil/GermanDefinitionGeneration-Distillation` for google/mt5-base

An [adapter](https://adapterhub.ml) for the `google/mt5-base` model that was trained on the [jfeil/GermanDefinitionGeneration-Distillation](https://huggingface.co/datasets/jfeil/GermanDefinitionGeneration-Distillation/) dataset.

This adapter was created for usage with the **[Adapters](https://github.com/Adapter-Hub/adapters)** library.

## Usage

First, install `adapters`:

```
pip install -U adapters
```

Now, the adapter can be loaded and activated like this:

```python
from adapters import AutoAdapterModel

model = AutoAdapterModel.from_pretrained("google/mt5-base")
adapter_name = model.load_adapter("jfeil/GermanDefinitionGeneration-Distillation", source="hf", set_active=True)
```

## Architecture & Training

TODO

## Optimal Inferencing Parameters

* num_beams	5
* repetition_penalty	1.5
* Prompt: `<EXAMPLE_SENTENCE> Was ist die Definition von <EXAMPLE_WORD>?`

## Evaluation results

Tested on Wiktionary test set:

|              | MeteorScore | ROUGE-L | BERTscore-F1 | MoverScore | Sentence Embedding Score |
|--------------|-------------|---------|--------------|------------|--------------------------|
| Wiktionary   | 0.175       | 0.187   | 0.751        | 0.568      | 0.478                    |
| Distillation | 0.103       | 0.109   | 0.675        | 0.536      | 0.369                    |

## Citation

TODO