Hugging Face's logo

OpenLLM-Ro
/
RoGemma-7b-Instruct-2024-06-28

Text Generation
Transformers
Safetensors
Romanian
gemma
conversational
Eval Results
text-generation-inference
Model card Files Community
1
RoGemma-7b-Instruct-2024-06-28 / README.md
mihaimasala's picture
mihaimasala
Update README.md
e711307 verified
preview code
|
raw
history blame
9.7 kB
---
license: cc-by-nc-4.0
language:
- ro
base_model:
- google/gemma-7b
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
RoGemma is a family of pretrained and fine-tuned generative text models for Romanian. This is the repository for the **instruct 7B model**. Links to other models can be found at the bottom of this page.
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
OpenLLM-Ro represents the first open-source effort to build a LLM specialized for Romanian. OpenLLM-Ro developed and publicly releases a collection of Romanian LLMs, both in the form of foundational model and instruct and chat variants.
- **Developed by:** OpenLLM-Ro
<!-- - **Funded by [optional]:** [More Information Needed] -->
<!-- - **Shared by [optional]:** [More Information Needed] -->
<!-- - **Model type:** [More Information Needed] -->
- **Language(s):** Romanian
- **License:** cc-by-nc-4.0
- **Finetuned from model:** [gemma-7b](https://huggingface.co/google/gemma-7b)
- **Trained using:** [RoAlpaca](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_alpaca), [RoAlpacaGPT4](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_alpaca_gpt4), [RoDolly](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_dolly), [RoSelfInstruct](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_selfinstruct_gpt4), [RoNoRobots](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_norobots), [RoOrca](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_orca), [RoCamel](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_camel)
### Model Sources
<!-- Provide the basic links for the model. -->
- **Repository:** https://github.com/OpenLLM-Ro/LLaMA-Factory
- **Paper:** https://arxiv.org/abs/2406.18266
## Intended Use
### Intended Use Cases
RoGemma is intented for research use in Romanian. Base models can be adapted for a variety of natural language tasks while instruction and chat tuned models are intended for assistant-like chat.
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
Use in any manner that violates the license, any applicable laws or regluations, use in languages other than Romanian.
## How to Get Started with the Model
Use the code below to get started with the model.
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("OpenLLM-Ro/RoGemma-7b-Instruct")
model = AutoModelForCausalLM.from_pretrained("OpenLLM-Ro/RoGemma-7b-Instruct")
instruction = "Ce jocuri de societate pot juca cu prietenii mei?"
chat = [
{"role": "user", "content": instruction},
]
prompt = tokenizer.apply_chat_template(chat, tokenize=False, system_message="")
inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt")
outputs = model.generate(input_ids=inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0]))
```
## Academic Benchmarks
<table>
<tbody>
<tr>
<td><strong>Model</strong></td>
<td><strong><center>Average</center></strong></td>
<td><strong><center>ARC</center></strong></td>
<td><strong><center>MMLU</center></strong></td>
<td><strong><center>Winogrande</center></strong></td>
<td><strong><center>Hellaswag</center></strong></td>
<td><strong><center>GSM8k</center></strong></td>
<td><strong><center>TruthfulQA</center></strong></td>
</tr>
<tr>
<td>gemma-1.1-7b-it</td><td><center>41.44</center></td><td><center>40.32</center></td><td><center>47.22</center></td><td><center>55.01</center></td><td><center>47.03</center></td><td><center>9.50</center></td><td><center>49.58</center></td>
</tr>
<tr>
<td><em>RoGemma-7b-Instruct</em></td><td><center><em><strong>53.42</strong></em></center></td><td><center><em><strong>52.44</strong></em></center></td><td><center><em><strong>54.44</strong></em></center></td><td><center><em><strong>69.36</strong></em></center></td><td><center><em><strong>61.96</strong></em></center></td><td><center><em><strong>31.06</strong></em></center></td><td><center><em><strong>51.23</strong></em></center></td>
</tr>
</tbody>
</table>
## Downstream tasks
<table>
<tbody>
<tr>
<td></td>
<td colspan="4"><center><strong>LaRoSeDa</strong></center></td>
<td colspan="4"><center><strong>WMT</strong></center></td>
</tr>
<tr>
<td></td>
<td colspan="2"><center><strong>Few-shot</strong></center></td>
<td colspan="2"><center><strong>Finetuned</strong></center></td>
<td colspan="2"><center><strong>Few-shot</strong></center></td>
<td colspan="2"><center><strong>Finetuned</strong></center></td>
</tr>
<tr>
<td><strong>Model</strong></td>
<td><center><strong>Binary<br>(Macro F1)</strong></center></td>
<td><center><strong>Multiclass<br>(Macro F1)</strong></center></td>
<td><center><strong>Binary<br>(Macro F1)</strong></center></td>
<td><center><strong>Multiclass<br>(Macro F1)</strong></center></td>
<td><center><strong>EN-RO<br>(Bleu)</strong></center></td>
<td><center><strong>RO-EN<br>(Bleu)</strong></center></td>
<td><center><strong>EN-RO<br>(Bleu)</strong></center></td>
<td><center><strong>RO-EN<br>(Bleu)</strong></center>
</tr>
<tr>
<td>gemma-1.1-7b-it</td><td><center>87.54</center></td><td><center>51.49</center></td><td><center>83.87</center></td><td><center>85.61</center></td><td><center>17.96</center></td><td><center><strong>27.74</strong></center></td><td><center>25.48</center></td><td><center>36.11</center></td>
</tr>
<tr>
<td><em>RoGemma-7b-Instruct</em></td><td><center><em><strong>97.87</strong></em></center></td><td><center><em><strong>65.71</strong></em></center></td><td><center><em><strong>98.43</strong></em></center></td><td><center><em><strong>87.18</strong></em></center></td><td><center><em><strong>27.91</strong></em></center></td><td><center><em>23.08</em></center></td><td><center><em><strong>27.99</strong></em></center></td><td><center><em><strong>39.51</strong></em></center></td>
</tr>
</tbody>
</table>
<table>
<tbody>
<tr>
<td></td>
<td colspan="4"><center><strong>XQuAD</strong></center></td>
<td colspan="4"><center><strong>STS</strong></center></td>
</tr>
<tr>
<td></td>
<td colspan="2"><center><strong>Few-shot</strong></center></td>
<td colspan="2"><center><strong>Finetuned</strong></center></td>
<td colspan="2"><center><strong>Few-shot</strong></center></td>
<td colspan="2"><center><strong>Finetuned</strong></center></td>
</tr>
<tr>
<td><strong>Model</strong></td>
<td><center><strong>(EM)</strong></center></td>
<td><center><strong>(F1)</strong></center></td>
<td><center><strong>(EM)</strong></center></td>
<td><center><strong>(F1)</strong></center></td>
<td><center><strong>(Spearman)</strong></center></td>
<td><center><strong>(Pearson)</strong></center></td>
<td><center><strong>(Spearman)</strong></center></td>
<td><center><strong>(Pearson)</strong></center></td>
</tr>
<tr>
<td>gemma-1.1-7b-it</td><td><center><strong>42.10</strong></center></td><td><center><strong>62.30</strong></center></td><td><center><strong>60.34</strong></center></td><td><center><strong>77.40</strong></center></td><td><center>49.10</center></td><td><center>50.23</center></td><td><center>83.43</center></td><td><center>83.65</center></td>
</tr>
<tr>
<td><em>RoGemma-7b-Instruct</em></td><td><center><em>17.75</em></center></td><td><center><em>28.11</em></center></td><td><center><em>52.02</em></center></td><td><center><em>68.43</em></center></td><td><center><em><strong>73.96</strong></em></center></td><td><center><em><strong>75.16</strong></em></center></td><td><center><em><strong>86.45</strong></em></center></td><td><center><em><strong>86.31</strong></em></center></td>
</tr>
</tbody>
</table>
## MT-Bench
<table>
<tbody>
<tr>
<td><strong>Model</strong></td>
<td><strong><center>Average</center></strong></td>
<td><strong><center>1st turn</center></strong></td>
<td><strong><center>2nd turn</center></strong></td>
<td><strong><center>Answers in Ro</center></strong></td>
</tr>
<tr>
<td>gemma-1.1-7b-it</td><td><center>4.83</center></td><td><center>5.11</center></td><td><center>4.55</center></td><td><center><strong>160/160</strong></center></td>
</tr>
<tr>
<td><em>RoGemma-7b-Instruct</em></td><td><center><em><strong>5.26</strong></em></center></td><td><center><em><strong>5.92</strong></em></center></td><td><center><em><strong>4.60</strong></em></center></td><td><center><em><strong>160/160</strong></em></center></td>
</tr>
</tbody>
</table>
## RoCulturaBench
<table>
<tbody>
<tr>
<td><strong>Model</strong></td>
<td><strong><center>Average</center></strong></td>
<td><strong><center>Answers in Ro</center></strong></td>
</tr>
<tr>
<td>gemma-1.1-7b-it</td><td><center><strong>3.38</strong></center></td><td><center><strong>100/100</strong></center></td>
</tr>
<tr>
<td><em>RoGemma-7b-Instruct</em></td><td><center><em>3.26</em></center></td><td><center><em><strong>100/100</strong></em></center></td>
</tr>
</tbody>
</table>
## RoGemma Model Family
| Model | Link |
|--------------------|:--------:|
|*RoGemma-7b-Instruct*| [link](https://huggingface.co/OpenLLM-Ro/RoGemma-7b-Instruct) |
## Citation
```
@misc{masala2024vorbecstiromanecsterecipetrain,
title={"Vorbe\c{s}ti Rom\^ane\c{s}te?" A Recipe to Train Powerful Romanian LLMs with English Instructions},
author={Mihai Masala and Denis C. Ilie-Ablachim and Alexandru Dima and Dragos Corlatescu and Miruna Zavelca and Ovio Olaru and Simina Terian-Dan and Andrei Terian-Dan and Marius Leordeanu and Horia Velicu and Marius Popescu and Mihai Dascalu and Traian Rebedea},
year={2024},
eprint={2406.18266},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2406.18266},
}
```
<!-- **APA:**
[More Information Needed] -->