|
--- |
|
datasets: |
|
- oscar-corpus/OSCAR-2301 |
|
- wikipedia |
|
- bjoernp/tagesschau-2018-2023 |
|
language: |
|
- en |
|
- de |
|
library_name: transformers |
|
pipeline_tag: text-generation |
|
license: llama2 |
|
--- |
|
# LAION LeoLM 70b: **L**inguistically **E**nhanced **O**pen **L**anguage **M**odel |
|
Meet LeoLM, the first open and commercially available German Foundation Language Model built on Llama-2. |
|
Our models extend Llama-2's capabilities into German through continued pretraining on a large corpus of German-language and mostly locality specific text. |
|
Thanks to a compute grant at HessianAI's new supercomputer **42**, we release a series foundation models trained with 8k context length |
|
under the [Llama-2 community license](https://huggingface.co/meta-llama/Llama-2-70b/raw/main/LICENSE.txt). Now, we're finally releasing the |
|
much anticipated `leo-hessianai-70b`, the largest model of this series based on `Llama-2-70b`. |
|
With this release, we hope to bring a new wave of opportunities to German open-source and commercial LLM research and accelerate adoption. |
|
Read our [blog post](https://laion.ai/blog/leo-lm/) or our paper (preprint coming soon) for more details! |
|
|
|
*A project by Björn Plüster and Christoph Schuhmann in collaboration with LAION and HessianAI.* |
|
|
|
|
|
## Model Details |
|
- **Finetuned from:** [meta-llama/Llama-2-70b-hf](https://huggingface.co/meta-llama/Llama-2-70b-hf) |
|
- **Model type:** Causal decoder-only transformer language model |
|
- **Language:** English and German |
|
- **License:** [LLAMA 2 COMMUNITY LICENSE AGREEMENT](https://huggingface.co/meta-llama/Llama-2-70b/raw/main/LICENSE.txt) |
|
- **Contact:** [LAION Discord](https://discord.com/invite/eq3cAMZtCC) or [Björn Plüster](mailto:bjoern.pl@outlook.de) |
|
|
|
|
|
## Use in 🤗Transformers |
|
First install direct dependencies: |
|
``` |
|
pip install transformers torch |
|
``` |
|
|
|
Then load the model in transformers. Note that this requires lots of VRAM and most-likely multiple devices. Use `load_in_8bit=True` or `load_in_4bit=True` |
|
to save some memory by using a quantized version. For more quantized versions, check out our models at TheBloke's page: (coming soon!) |
|
```python |
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
import torch |
|
|
|
model = AutoModelForCausalLM.from_pretrained( |
|
model="LeoLM/leo-hessianai-70b", |
|
device_map="auto", |
|
torch_dtype=torch.bfloat16, |
|
use_flash_attention_2=False # Set to true to use FA2. Requires `pip install flash-attn` |
|
) |
|
``` |
|
|
|
## Training parameters |
|
 |
|
|
|
|
|
## Benchmarks |
|
 |
|
 |
