138 lines
8.2 KiB
Markdown
138 lines
8.2 KiB
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
- es
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
---
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
# Occiglot-7B-FR-EN-Instruct
|
||
|
|
|
||
|
|
> A [polyglot](https://en.wikipedia.org/wiki/Multilingualism#In_individuals) language model for the [Occident](https://en.wikipedia.org/wiki/Occident).
|
||
|
|
>
|
||
|
|
|
||
|
|
**Occiglot-7B-FR-EN-Instruct** is a the instruct version of [occiglot-7b-fr-en](https://huggingface.co/occiglot/occiglot-7b-fr-en), a generative language model with 7B parameters supporting the Spanish and English and trained by the [Occiglot Research Collective](https://occiglot.github.io/occiglot/).
|
||
|
|
It was trained on 160M tokens of additional multilingual and code instructions.
|
||
|
|
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
|
||
|
|
|
||
|
|
- **Instruction tuned from:** [occiglot-7b-fr-en](https://huggingface.co/occiglot/occiglot-7b-fr-en)
|
||
|
|
- **Model type:** Causal decoder-only transformer language model
|
||
|
|
- **Languages:** English, Spanish, 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)
|
||
|
|
|
||
|
|
### 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-es-en-instruct")
|
||
|
|
>>> model = MistralForCausalLM.from_pretrained('occiglot/occiglot-7b-es-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": "qui est le président français ?"},
|
||
|
|
>>> ]
|
||
|
|
>>> 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]):])
|
||
|
|
'Le président français est Emmanuel Macron.'
|
||
|
|
```
|
||
|
|
|
||
|
|
## Dataset
|
||
|
|
|
||
|
|
The training data was split evenly amongst French and English based on the total number of tokens.
|
||
|
|
|
||
|
|
**English and Code**
|
||
|
|
- [Open-Hermes-2B](https://huggingface.co/datasets/teknium/OpenHermes-2.5)
|
||
|
|
|
||
|
|
|
||
|
|
**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)
|
||
|
|
|
||
|
|
## 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>
|
||
|
|
|
||
|
|
|
||
|
|
### English
|
||
|
|
|
||
|
|
| | arc_challenge | belebele | hellaswag | mmlu | truthfulqa | avg |
|
||
|
|
|:-------------------------------------|----------------:|-----------:|------------:|---------:|-------------:|---------:|
|
||
|
|
| occiglot/occiglot-7b-eu5 | 0.530717 | 0.726667 | 0.789882 | 0.531904 | 0.403678 | 0.59657 |
|
||
|
|
| occiglot/occiglot-7b-eu5-instruct | 0.558874 | 0.746667 | 0.799841 | 0.535109 | 0.449034 | 0.617905 |
|
||
|
|
| occiglot/occiglot-7b-de-en | 0.556314 | 0.791111 | 0.803824 | 0.568438 | 0.423251 | 0.628587 |
|
||
|
|
| occiglot/occiglot-7b-de-en-instruct | 0.604096 | 0.812222 | 0.80004 | 0.570574 | 0.493807 | 0.656148 |
|
||
|
|
| LeoLM/leo-mistral-hessianai-7b | 0.522184 | 0.736667 | 0.777833 | 0.538812 | 0.429248 | 0.600949 |
|
||
|
|
| mistralai/Mistral-7B-v0.1 | 0.612628 | 0.844444 | 0.834097 | 0.624555 | 0.426201 | 0.668385 |
|
||
|
|
| mistralai/Mistral-7B-Instruct-v0.2 | 0.637372 | 0.824444 | 0.846345 | 0.59201 | 0.668116 | 0.713657 |
|
||
|
|
|
||
|
|
### French
|
||
|
|
|
||
|
|
| | arc_challenge_fr | belebele_fr | hellaswag_fr | mmlu_fr | truthfulqa_fr | avg |
|
||
|
|
|:-------------------------------------|-------------------:|--------------:|---------------:|----------:|----------------:|---------:|
|
||
|
|
| occiglot/occiglot-7b-eu5 | 0.506416 | 0.675556 | 0.712358 | 0.495684 | 0.23507 | 0.525017 |
|
||
|
|
| occiglot/occiglot-7b-eu5-instruct | 0.541488 | 0.7 | 0.724245 | 0.499122 | 0.306226 | 0.554216 |
|
||
|
|
| occiglot/occiglot-7b-fr-en | 0.532934 | 0.706667 | 0.718891 | 0.51333 | 0.242694 | 0.542903 |
|
||
|
|
| occiglot/occiglot-7b-fr-en-instruct | 0.542344 | 0.752222 | 0.72553 | 0.52051 | 0.29479 | 0.567079 |
|
||
|
|
| OpenLLM-France/Claire-Mistral-7B-0.1 | 0.486741 | 0.694444 | 0.642964 | 0.479566 | 0.271919 | 0.515127 |
|
||
|
|
| mistralai/Mistral-7B-v0.1 | 0.525235 | 0.776667 | 0.66481 | 0.543121 | 0.280813 | 0.558129 |
|
||
|
|
| mistralai/Mistral-7B-Instruct-v0.2 | 0.551754 | 0.758889 | 0.67916 | 0.506837 | 0.382465 | 0.575821 |
|
||
|
|
|
||
|
|
</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
|
||
|
|
- https://huggingface.co/NikolayKozloff/occiglot-7b-fr-en-GGUF
|