|
--- |
|
license: apache-2.0 |
|
language: |
|
- en |
|
- es |
|
- de |
|
- fr |
|
- it |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
![image/png](https://huggingface.co/datasets/malteos/images/resolve/main/occiglot.medium.png) |
|
|
|
# Occiglot-7B-EU5 |
|
|
|
> A [polyglot](https://en.wikipedia.org/wiki/Multilingualism#In_individuals) language model for the [Occident](https://en.wikipedia.org/wiki/Occident). |
|
> |
|
|
|
**Occiglot-7B-EU5-Instruct** is a the instruct version of [occiglot-7b-eu5](https://huggingface.co/occiglot/occiglot-7b-eu5/edit/main/README.md), a generative language model with 7B parameters supporting the top-5 EU languages (English, Spanish, French, German, and Italian) and trained by the [Occiglot Research Collective](https://occiglot.github.io/occiglot/). |
|
It is based on [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) and trained on 400M tokens of additional multilingual and code instructions with a block size of 8,192 tokens per sample. |
|
Note that the model was not safety aligned and might generate problematic outputs. |
|
|
|
This is the first release of an ongoing open research project for multilingual language models. |
|
If you want to train a model for your own language or are working on evaluations, please contact us or join our [Discord server](https://discord.gg/wUpvYs4XvM). **We are open for collaborations!** |
|
|
|
|
|
### Model details |
|
|
|
- **Continued-pretraining from:** [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) |
|
- **Model type:** Causal decoder-only transformer language model |
|
- **Languages:** English, Spanish, French, German, Italian, and code. |
|
- **License:** [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0.html) |
|
- **Compute resources:** [DFKI cluster](https://www.dfki.de/en/web) |
|
- **Contributors:** Manuel Brack, Patrick Schramowski, Pedro Ortiz, Malte Ostendorff, Fabio Barth, Georg Rehm, Kristian Kersting |
|
- **Research labs:** [Occiglot](https://occiglot.github.io/occiglot/) with support from [SAINT](https://www.dfki.de/en/web/research/research-departments/foundations-of-systems-ai) and [SLT](https://www.dfki.de/en/web/research/research-departments/speech-and-language-technology) |
|
- **Contact:** [Discord](https://discord.gg/wUpvYs4XvM) [hello@occiglot.org](mailto:hello@occiglot.org) |
|
|
|
### How to use |
|
|
|
The model was trained using the chatml instruction template. You can use the transformers chat template feature for interaction. |
|
Since the generation relies on some randomness, we |
|
set a seed for reproducibility: |
|
|
|
```python |
|
>>> from transformers import AutoTokenizer, MistralForCausalLM, set_seed |
|
>>> tokenizer = AutoTokenizer.from_pretrained("occiglot/occiglot-7b-de-en-instruct") |
|
>>> model = MistralForCausalLM.from_pretrained('occiglot/occiglot-7b-de-en-instruct') # You may want to use bfloat16 and/or move to GPU here |
|
>>> set_seed(42) |
|
>>> messages = [ |
|
>>> {"role": "system", 'content': 'You are a helpful assistant. Please give short and concise answers.'}, |
|
>>> {"role": "user", "content": "Wer ist der deutsche Bundeskanzler?"}, |
|
>>> ] |
|
>>> tokenized_chat = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_dict=False, return_tensors='pt',) |
|
>>> set_seed(42) |
|
>>> outputs = model.generate(tokenized_chat.to('cuda'), max_new_tokens=200,) |
|
>>> tokenizer.decode(out[0][len(tokenized_chat[0]):]) |
|
'Der deutsche Bundeskanzler ist Olaf Scholz.' |
|
``` |
|
|
|
## Dataset |
|
|
|
The training data was split evenly amongst the 5 languages based on the total number of tokens. We would like to thank Disco Research and Björn Plüster for making their dataset available to us. |
|
|
|
**English and Code** |
|
- [Open-Hermes-2B](https://huggingface.co/datasets/teknium/OpenHermes-2.5) |
|
|
|
**German** |
|
- [DiscoLM German Dataset](https://huggingface.co/DiscoResearch) |
|
- [OASST-2](https://huggingface.co/datasets/OpenAssistant/oasst2) (German subset) |
|
- [Aya-Dataset](https://huggingface.co/datasets/CohereForAI/aya_dataset) (German subset) |
|
|
|
**Spanish** |
|
- [Mentor-ES](https://huggingface.co/datasets/projecte-aina/MentorES) |
|
- [Squad-es](https://huggingface.co/datasets/squad_es) |
|
- [OASST-2](https://huggingface.co/datasets/OpenAssistant/oasst2) (Spanish subset) |
|
- [Aya-Dataset](https://huggingface.co/datasets/CohereForAI/aya_dataset) (Spanish subset) |
|
|
|
**French** |
|
- [Bactrian-X](https://huggingface.co/datasets/MBZUAI/Bactrian-X) (French subset) |
|
- [AI-Society Translated](https://huggingface.co/datasets/camel-ai/ai_society_translated) (French subset) |
|
- [GT-Dorimiti](https://huggingface.co/datasets/Gt-Doremiti/gt-doremiti-instructions) |
|
- [OASST-2](https://huggingface.co/datasets/OpenAssistant/oasst2) (French subset) |
|
- [Aya-Dataset](https://huggingface.co/datasets/CohereForAI/aya_dataset) (French subset) |
|
|
|
**Italian** |
|
- [Quora-IT-Baize](https://huggingface.co/datasets/andreabac3/Quora-Italian-Fauno-Baize) |
|
- [Stackoverflow-IT-Vaize](https://huggingface.co/datasets/andreabac3/StackOverflow-Italian-Fauno-Baize) |
|
- [Camoscio](https://huggingface.co/datasets/teelinsan/camoscio_cleaned) |
|
- [OASST-2](https://huggingface.co/datasets/OpenAssistant/oasst2) (Italian subset) |
|
- [Aya-Dataset](https://huggingface.co/datasets/CohereForAI/aya_dataset) (Italian subset) |
|
|
|
## Training settings |
|
|
|
- Full instruction fine-tuning on 8xH100. |
|
- 0.6 - 4 training epochs (depending on dataset sampling). |
|
- Framework: [axolotl](https://github.com/OpenAccess-AI-Collective/axolotl) |
|
- Precision: bf16 |
|
- Optimizer: AdamW |
|
- Global batch size: 128 (with 8192 context length) |
|
- Cosine Annealing with Warmup |
|
|
|
|
|
## Tokenizer |
|
|
|
Tokenizer is unchanged from [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1). |
|
|
|
## Evaluation |
|
|
|
Preliminary evaluation results can be found below. |
|
Please note that the non-English results are based on partially machine-translated datasets and English prompts ([Belebele](https://huggingface.co/datasets/facebook/belebele) and [Okapi framework](https://github.com/nlp-uoregon/Okapi)) and thus should be interpreted with caution, e.g., biased towards English model performance. |
|
Currently, we are working on more suitable benchmarks for Spanish, French, German, and Italian. |
|
|
|
<details> |
|
<summary>Evaluation results</summary> |
|
|
|
|
|
</details> |
|
|
|
## Acknowledgements |
|
|
|
The pre-trained model training was supported by a compute grant at the [42 supercomputer](https://hessian.ai/) which is a central component in the development of [hessian AI](https://hessian.ai/), the [AI Innovation Lab](https://hessian.ai/infrastructure/ai-innovationlab/) (funded by the [Hessian Ministry of Higher Education, Research and the Art (HMWK)](https://wissenschaft.hessen.de) & the [Hessian Ministry of the Interior, for Security and Homeland Security (HMinD)](https://innen.hessen.de)) and the [AI Service Centers](https://hessian.ai/infrastructure/ai-service-centre/) (funded by the [German Federal Ministry for Economic Affairs and Climate Action (BMWK)](https://www.bmwk.de/Navigation/EN/Home/home.html)). |
|
The curation of the training data is partially funded by the [German Federal Ministry for Economic Affairs and Climate Action (BMWK)](https://www.bmwk.de/Navigation/EN/Home/home.html) |
|
through the project [OpenGPT-X](https://opengpt-x.de/en/) (project no. 68GX21007D). |
|
|
|
|
|
## License |
|
|
|
[Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0.html) |
|
|
|
## See also |
|
|
|
- https://huggingface.co/collections/occiglot/occiglot-eu5-7b-v01-65dbed502a6348b052695e01 |
|
|
|
|