Zion_Alpha_Instruction_Tuned

<div align="center"> <b style="font-size: 40px;">Zion_Alpha_Instruction_Tuned</b> </div> <img src="https://i.imgur.com/e1LEQ18.png" alt="Zion_Alpha_Instruction_Tuned" style="width: 50%; min-width: 400px; display: block; margin: auto;">

Model Details

Zion_Alpha is the first REAL Hebrew model in the world. This version WAS fine tuned for tasks. I did the finetune using SOTA techniques and using my insights from years of underwater basket weaving. If you wanna offer me a job, just add me on Facebook.

Future Plans

I plan to perform a SLERP merge with one of my other fine-tuned models, which has a bit more knowledge about Israeli topics. Additionally, I might create a larger model using MergeKit, but we'll see how it goes.

Looking for Sponsors

Since all my work is done on-premises, I am constrained by my current hardware. I would greatly appreciate any support in acquiring an A6000, which would enable me to train significantly larger models much faster.

Papers?

Maybe. We'll see. No promises here 🤓

Contact Details

I'm not great at self-marketing (to say the least) and don't have any social media accounts. If you'd like to reach out to me, you can email me at [email protected]. Please note that this email might receive more messages than I can handle, so I apologize in advance if I can't respond to everyone.

Versions and QUANTS

• Base model: FP16
• Instruction tuned: FP16 | GGUF

Model architecture

Based on Mistral 7B. I didn't even bother to alter the tokenizer.

The recommended prompt setting is Debug-deterministic:

temperature: 1
top_p: 1
top_k: 1
typical_p: 1
min_p: 1
repetition_penalty: 1

The recommended instruction template is Mistral:

{%- for message in messages %}
    {%- if message['role'] == 'system' -%}
        {{- message['content'] -}}
    {%- else -%}
        {%- if message['role'] == 'user' -%}
            {{-'[INST] ' + message['content'].rstrip() + ' [/INST]'-}}
        {%- else -%}
            {{-'' + message['content'] + '</s>' -}}
        {%- endif -%}
    {%- endif -%}
{%- endfor -%}
{%- if add_generation_prompt -%}
    {{-''-}}
{%- endif -%}

English to hebrew example:

<div align="center"> <b style="font-size: 40px;">Zion_Alpha English to Hebrew example</b> </div> <img src="https://i.imgur.com/JnTuawF.png" alt="Zion_Alpha" style="width: 40%; min-width: 600px; display: block; margin: auto;">

English to hebrew example:

<div align="center"> <b style="font-size: 40px;">Zion_Alpha Hebrew to English example</b> </div> <img src="https://i.imgur.com/Wm2igLJ.png" alt="Zion_Alpha" style="width: 40%; min-width: 600px; display: block; margin: auto;"> <div align="center"> <b style="font-size: 30px;">Unscripted video: live zero shot demonstration at story writing capabilities in Hebrew</b>

</div> <div align="center"> <b style="font-size: 30px;">Zion_Alpha VS Mistral 'Hebrew' Live & unscripted in real time</b>

</div> <div align="center"> <b style="font-size: 30px;">Zion_Alpha VS Mistral 'Hebrew' Live & unscripted in real time Long text translation</b>

</div>

History

The model was originally trained about 2 month after Mistral (v0.1) was released. As of 04 June 2024, Zion_Alpha got the Highest SNLI score in the world among open source models in Hebrew, surpassing most of the models by a huge margin. (84.05 score)

<img src="https://huggingface.co/SicariusSicariiStuff/Zion_Alpha/resolve/main/Images/Zion_Alpha.png" alt="Zion_Alpha SNLI Score" style="width: 100%; min-width: 700px; display: block; margin: auto;">

---

Citation Information

@llm{Zion_Alpha_Instruction_Tuned,
  author = {SicariusSicariiStuff},
  title = {Zion_Alpha_Instruction_Tuned},
  year = {2024},
  publisher = {Hugging Face},
  url = {https://huggingface.co/SicariusSicariiStuff/Zion_Alpha_Instruction_Tuned}
}

---

Support

<img src="https://i.imgur.com/0lHHN95.png" alt="GPUs too expensive" style="width: 10%; min-width: 100px; display: block; margin: left;">

• My Ko-fi page ALL donations will go for research resources and compute, every bit counts 🙏🏻
• My Patreon ALL donations will go for research resources and compute, every bit counts 🙏🏻