Shaltiel's picture
Update README.md
9ed3346 verified
---
license: apache-2.0
pipeline_tag: text-generation
language:
- en
- he
tags:
- instruction-tuned
base_model: dicta-il/dictalm2.0
inference: false
---
[<img src="https://i.ibb.co/5Lbwyr1/dicta-logo.jpg" width="300px"/>](https://dicta.org.il)
# Adapting LLMs to Hebrew: Unveiling DictaLM 2.0 with Enhanced Vocabulary and Instruction Capabilities
The DictaLM-2.0-Instruct Large Language Model (LLM) is an instruct fine-tuned version of the [DictaLM-2.0](https://huggingface.co/dicta-il/dictalm2.0) generative model using a variety of conversation datasets.
For full details of this model please read our [release blog post](https://dicta.org.il/dicta-lm) or the [technical report](https://arxiv.org/abs/2407.07080).
This is the instruct-tuned model designed for chat in the GGUF format for use with [LM Studio](https://lmstudio.ai/) or [llama.cpp](https://github.com/ggerganov/llama.cpp). You can try the model out on a live demo [here](https://huggingface.co/spaces/dicta-il/dictalm2.0-instruct-demo).
There are two versions available - float16 precision (`*.F16.gguf`) and 4-bit quantized precision (`*.Q4_K_M.gguf`).
You can view and access the full collection of base/instruct unquantized/quantized versions of `DictaLM-2.0` [here](https://huggingface.co/collections/dicta-il/dicta-lm-20-collection-661bbda397df671e4a430c27).
## Instruction format
In order to leverage instruction fine-tuning, your prompt should be surrounded by `[INST]` and `[/INST]` tokens followed by a line break. The very first instruction should begin with a begin of sentence id. The next instructions should not. The assistant generation will be ended by the end-of-sentence token id.
E.g.
```
text = """<s>[INST] 讗讬讝讛 专讜讟讘 讗讛讜讘 注诇讬讱? [/INST]
讟讜讘, 讗谞讬 讚讬 诪讞讘讘 讻诪讛 讟讬驻讜转 诪讬抓 诇讬诪讜谉 住讞讜讟 讟专讬. 讝讛 诪讜住讬祝 讘讚讬讜拽 讗转 讛讻诪讜转 讛谞讻讜谞讛 砖诇 讟注诐 讞诪爪诪抓 诇讻诇 诪讛 砖讗谞讬 诪讘砖诇 讘诪讟讘讞!</s>[INST] 讛讗诐 讬砖 诇讱 诪转讻讜谞讬诐 诇诪讬讜谞讝? [/INST]"
```
This format is available as a [chat template](https://huggingface.co/docs/transformers/main/chat_templating) via the `apply_chat_template()` method:
## Using with LM Studio
When using with LM Studio, just search the hub for "dictalm2.0-instruct-GGUF", and the model in both precisions should appear.
Make sure to set the chat template correctly - initialize from the `mistral-instruct` template, and add a `\n` in the suffix box, like here:
<img src="https://i.ibb.co/D9MVgK2/lmstudio-dlm-template.png" width="400px" />
In addition, the model doesn't support any system prompt, so make sure to remove the system prompt as well.
## Model Architecture
DictaLM-2.0-Instruct follows the [Zephyr-7B-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) recipe for fine-tuning an instruct model, with an extended instruct dataset for Hebrew.
## Limitations
The DictaLM 2.0 Instruct model is a demonstration that the base model can be fine-tuned to achieve compelling performance.
It does not have any moderation mechanisms. We're looking forward to engaging with the community on ways to
make the model finely respect guardrails, allowing for deployment in environments requiring moderated outputs.
## Citation
If you use this model, please cite:
```bibtex
@misc{shmidman2024adaptingllmshebrewunveiling,
title={Adapting LLMs to Hebrew: Unveiling DictaLM 2.0 with Enhanced Vocabulary and Instruction Capabilities},
author={Shaltiel Shmidman and Avi Shmidman and Amir DN Cohen and Moshe Koppel},
year={2024},
eprint={2407.07080},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2407.07080},
}
```