vaiton/minerva

vaiton/ minerva:latest

1,440 Downloads Updated 2 years ago

The first family of Large Language Models pretrained from scratch on Italian!

ollama run vaiton/minerva

curl http://localhost:11434/api/chat \
  -d '{
    "model": "vaiton/minerva",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='vaiton/minerva',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'vaiton/minerva',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 2 years ago

2 years ago

68a06aa7974d · 3.1GB ·

model

archllama

parameters2.89B

quantizationQ8_0

3.1GB

Readme

Minerva is the first family of LLMs pretrained from scratch on Italian developed by Sapienza NLP in collaboration with Future Artificial Intelligence Research (FAIR) and CINECA.

The Minerva models are truly-open (data and model) Italian-English LLMs, with approximately half of the pretraining data composed of Italian text.

Tag	HuggingFace
`3b-base_Q8_0`	https://huggingface.co/VaiTon/Minerva-3B-base-v1.0-Q8_0-GGUF
`3b-base_Q5_K_M`	https://huggingface.co/VaiTon/Minerva-3B-base-v1.0-Q5_K_M-GGUF
`3b-instruct-Q5_K_M`	https://huggingface.co/NikolayKozloff/Minerva-3B-Instruct-v1.0-Q5_K_M-GGUF
`1b-base-Q8_0`	https://huggingface.co/VaiTon/Minerva-1B-base-v1.0-Q8_0-GGUF