Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) RoLlama3-8b-Instruct - GGUF - Model creator: https://huggingface.co/OpenLLM-Ro/ - Original model: https://huggingface.co/OpenLLM-Ro/RoLlama3-8b-Instruct/ | Name | Quant method | Size | | ---- | ---- | ---- | | [RoLlama3-8b-Instruct.Q2_K.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q2_K.gguf) | Q2_K | 2.96GB | | [RoLlama3-8b-Instruct.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.IQ3_XS.gguf) | IQ3_XS | 3.28GB | | [RoLlama3-8b-Instruct.IQ3_S.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.IQ3_S.gguf) | IQ3_S | 3.43GB | | [RoLlama3-8b-Instruct.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q3_K_S.gguf) | Q3_K_S | 3.41GB | | [RoLlama3-8b-Instruct.IQ3_M.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.IQ3_M.gguf) | IQ3_M | 3.52GB | | [RoLlama3-8b-Instruct.Q3_K.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q3_K.gguf) | Q3_K | 3.74GB | | [RoLlama3-8b-Instruct.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q3_K_M.gguf) | Q3_K_M | 3.74GB | | [RoLlama3-8b-Instruct.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q3_K_L.gguf) | Q3_K_L | 4.03GB | | [RoLlama3-8b-Instruct.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.IQ4_XS.gguf) | IQ4_XS | 4.18GB | | [RoLlama3-8b-Instruct.Q4_0.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q4_0.gguf) | Q4_0 | 4.34GB | | [RoLlama3-8b-Instruct.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.IQ4_NL.gguf) | IQ4_NL | 4.38GB | | [RoLlama3-8b-Instruct.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q4_K_S.gguf) | Q4_K_S | 4.37GB | | [RoLlama3-8b-Instruct.Q4_K.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q4_K.gguf) | Q4_K | 4.58GB | | [RoLlama3-8b-Instruct.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q4_K_M.gguf) | Q4_K_M | 4.58GB | | [RoLlama3-8b-Instruct.Q4_1.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q4_1.gguf) | Q4_1 | 4.78GB | | [RoLlama3-8b-Instruct.Q5_0.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q5_0.gguf) | Q5_0 | 5.21GB | | [RoLlama3-8b-Instruct.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q5_K_S.gguf) | Q5_K_S | 5.21GB | | [RoLlama3-8b-Instruct.Q5_K.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q5_K.gguf) | Q5_K | 5.34GB | | [RoLlama3-8b-Instruct.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q5_K_M.gguf) | Q5_K_M | 5.34GB | | [RoLlama3-8b-Instruct.Q5_1.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q5_1.gguf) | Q5_1 | 5.65GB | | [RoLlama3-8b-Instruct.Q6_K.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q6_K.gguf) | Q6_K | 6.14GB | | [RoLlama3-8b-Instruct.Q8_0.gguf](https://huggingface.co/RichardErkhov/OpenLLM-Ro_-_RoLlama3-8b-Instruct-gguf/blob/main/RoLlama3-8b-Instruct.Q8_0.gguf) | Q8_0 | 7.95GB | Original model description: --- license: cc-by-nc-4.0 language: - ro --- # Model Card for Model ID *Built with Meta Llama 3* RoLlama3 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 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 - **Language(s):** Romanian - **License:** cc-by-nc-4.0 - **Finetuned from model:** [Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) ### Model Sources - **Repository:** https://github.com/OpenLLM-Ro/llama-recipes - **Paper:** https://arxiv.org/abs/2406.18266 ## Intended Use ### Intended Use Cases RoLlama3 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 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/RoLlama3-8b-Instruct") model = AutoModelForCausalLM.from_pretrained("OpenLLM-Ro/RoLlama3-8b-Instruct") instruction = "Ce jocuri de societate pot juca cu prietenii mei?" chat = [ {"role": "system", "content": "Ești un asistent folositor, respectuos și onest. Încearcă să ajuți cât mai mult prin informațiile oferite, excluzând răspunsuri toxice, rasiste, sexiste, periculoase și ilegale."}, {"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])) ``` ## Benchmarks | Model | Average | ARC | MMLU |Winogrande|HellaSwag | GSM8k |TruthfulQA| |--------------------|:--------:|:--------:|:--------:|:--------:|:--------:|:--------:|:--------:| | Llama-3-8B-Instruct| 50.15 | 43.73 | 49.02 | 59.35 | 53.16 | **44.12** | 51.52 | | *RoLlama3-8b-Instruct* | ***50.61*** | ***44.66*** | ***52.19*** | ***67.58*** | ***57.65*** | *30.20* | ***51.39*** | ## MT-Bench | Model | Average | 1st turn | 2nd turn | Answers in Ro | |--------------------|:--------:|:--------:|:--------:|:--------:| | Llama-3-8B-Instruct | **5.92** | **6.36** | **5.49** | 158 / 160 | *RoLlama3-8b-Instruct*| *5.28* |*6.10*| *4.45* | ***160 / 160*** | ## RoCulturaBench | Model | Score | Answers in Ro| |--------------------|:--------:|:--------:| | Llama-3-8B-Instruct | **4.61** | **100 / 100** | | *RoLlama3-8b-Instruct*| *3.83*| ***100 / 100*** | ## RoLlama3 Model Family | Model | Link | |--------------------|:--------:| |*RoLlama3-8b-Instruct*| [link](https://huggingface.co/OpenLLM-Ro/RoLlama3-8b-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}, } ```