RoMistral 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.
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.
RoMistral 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.
Use in any manner that violates the license, any applicable laws or regluations, use in languages other than Romanian.
Use the code below to get started with the model.
1from transformers import AutoTokenizer, AutoModelForCausalLM
2
3tokenizer = AutoTokenizer.from_pretrained("OpenLLM-Ro/RoMistral-7b-Instruct")
4model = AutoModelForCausalLM.from_pretrained("OpenLLM-Ro/RoMistral-7b-Instruct")
5
6instruction = "Ce jocuri de societate pot juca cu prietenii mei?"
7chat = [
8 {"role": "user", "content": instruction},
9 ]
10prompt = tokenizer.apply_chat_template(chat, tokenize=False, system_message="")
11
12inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt")
13outputs = model.generate(input_ids=inputs, max_new_tokens=128)
14print(tokenizer.decode(outputs[0]))
