File size: 1,541 Bytes
148c76a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f5e3761
148c76a
 
 
 
 
 
f5e3761
148c76a
 
 
 
 
 
 
 
 
f5e3761
148c76a
c7cf5a6
 
 
 
 
 
148c76a
 
 
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

---
library_name: transformers
pipeline_tag: text-generation
language:
- multilingual
tags:
- generation
- question answering
- instruction tuning
datasets:
- MBZUAI/Bactrian-X
license: cc-by-nc-4.0
---

###  Model Description

This HF repository hosts instruction fine-tuned multilingual BLOOM model using the parallel instruction dataset called Bactrain-X in 52 languages. 
We progressively add a language during instruction fine-tuning at each time, and train 52 models in total. Then, we evaluate those models in three multilingual benchmarks. 

Please refer to [our paper](https://arxiv.org/abs/2404.04850) for more details.

* Base model: [BLOOM 7B1](https://huggingface.co/bigscience/bloom-7b1)
* Instruction languages: English, Chinese, Afrikaans, Arabic, Azerbaijani, Bengali, Czech, German, Spanish
* Instruction language codes: en, zh, af, ar, az, bn, cs, de, es
* Training method: full-parameter fine-tuning.

### Usage
The model checkpoint should be loaded using `transformers` library.

```python
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("MaLA-LM/lucky52-bloom-7b1-no-9")
model = AutoModelForCausalLM.from_pretrained("MaLA-LM/lucky52-bloom-7b1-no-9")
```

### Citation
```
@inproceedings{ji2025lucky52,
      title={How Many Languages Make Good Multilingual Instruction Tuning? A Case Study on BLOOM}, 
      author={Shaoxiong Ji and Pinzhen Chen},
      year={2025},
      booktitle={Proceedings of COLING},
      url={https://arxiv.org/abs/2404.04850}, 
}
```