Quantization made by Richard Erkhov.

[Github](https://github.com/RichardErkhov)

[Discord](https://discord.gg/pvy7H8DZMG)

[Request more models](https://github.com/RichardErkhov/quant_request)


LLaMAntino-2-7b-hf-ITA - GGUF
- Model creator: https://huggingface.co/swap-uniba/
- Original model: https://huggingface.co/swap-uniba/LLaMAntino-2-7b-hf-ITA/


| Name | Quant method | Size |
| ---- | ---- | ---- |
| [LLaMAntino-2-7b-hf-ITA.Q2_K.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q2_K.gguf) | Q2_K | 2.36GB |
| [LLaMAntino-2-7b-hf-ITA.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.IQ3_XS.gguf) | IQ3_XS | 2.6GB |
| [LLaMAntino-2-7b-hf-ITA.IQ3_S.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.IQ3_S.gguf) | IQ3_S | 2.75GB |
| [LLaMAntino-2-7b-hf-ITA.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q3_K_S.gguf) | Q3_K_S | 2.75GB |
| [LLaMAntino-2-7b-hf-ITA.IQ3_M.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.IQ3_M.gguf) | IQ3_M | 2.9GB |
| [LLaMAntino-2-7b-hf-ITA.Q3_K.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q3_K.gguf) | Q3_K | 3.07GB |
| [LLaMAntino-2-7b-hf-ITA.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q3_K_M.gguf) | Q3_K_M | 3.07GB |
| [LLaMAntino-2-7b-hf-ITA.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q3_K_L.gguf) | Q3_K_L | 3.35GB |
| [LLaMAntino-2-7b-hf-ITA.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.IQ4_XS.gguf) | IQ4_XS | 3.4GB |
| [LLaMAntino-2-7b-hf-ITA.Q4_0.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q4_0.gguf) | Q4_0 | 3.56GB |
| [LLaMAntino-2-7b-hf-ITA.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.IQ4_NL.gguf) | IQ4_NL | 3.58GB |
| [LLaMAntino-2-7b-hf-ITA.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q4_K_S.gguf) | Q4_K_S | 3.59GB |
| [LLaMAntino-2-7b-hf-ITA.Q4_K.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q4_K.gguf) | Q4_K | 3.8GB |
| [LLaMAntino-2-7b-hf-ITA.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q4_K_M.gguf) | Q4_K_M | 3.8GB |
| [LLaMAntino-2-7b-hf-ITA.Q4_1.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q4_1.gguf) | Q4_1 | 3.95GB |
| [LLaMAntino-2-7b-hf-ITA.Q5_0.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q5_0.gguf) | Q5_0 | 4.33GB |
| [LLaMAntino-2-7b-hf-ITA.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q5_K_S.gguf) | Q5_K_S | 4.33GB |
| [LLaMAntino-2-7b-hf-ITA.Q5_K.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q5_K.gguf) | Q5_K | 4.45GB |
| [LLaMAntino-2-7b-hf-ITA.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q5_K_M.gguf) | Q5_K_M | 4.45GB |
| [LLaMAntino-2-7b-hf-ITA.Q5_1.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q5_1.gguf) | Q5_1 | 4.72GB |
| [LLaMAntino-2-7b-hf-ITA.Q6_K.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q6_K.gguf) | Q6_K | 5.15GB |
| [LLaMAntino-2-7b-hf-ITA.Q8_0.gguf](https://huggingface.co/RichardErkhov/swap-uniba_-_LLaMAntino-2-7b-hf-ITA-gguf/blob/main/LLaMAntino-2-7b-hf-ITA.Q8_0.gguf) | Q8_0 | 6.67GB |




Original model description:
---
license: llama2
language:
- it
tags:
- text-generation-inference
---
# Model Card for LLaMAntino-2-7b-ITA
*Last Update: 22/01/2024*<br>

## Model description

<!-- Provide a quick summary of what the model is/does. -->

**LLaMAntino-2-7b** is a *Large Language Model (LLM)* that is an italian-adapted **LLaMA 2**. 
This model aims to provide Italian NLP researchers with a base model for natural language generation tasks.

The model was trained using *QLora* and using as training data [clean_mc4_it medium](https://huggingface.co/datasets/gsarti/clean_mc4_it/viewer/medium). 
If you are interested in more details regarding the training procedure, you can find the code we used at the following link:
- **Repository:** https://github.com/swapUniba/LLaMAntino

**NOTICE**: the code has not been released yet, we apologize for the delay, it will be available asap!

- **Developed by:** Pierpaolo Basile, Elio Musacchio, Marco Polignano, Lucia Siciliani, Giuseppe Fiameni, Giovanni Semeraro
- **Funded by:** PNRR project FAIR - Future AI Research
- **Compute infrastructure:** [Leonardo](https://www.hpc.cineca.it/systems/hardware/leonardo/) supercomputer
- **Model type:** LLaMA 2
- **Language(s) (NLP):** Italian
- **License:** Llama 2 Community License 
- **Finetuned from model:** [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf)

## How to Get Started with the Model

Below you can find an example of model usage:

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "swap-uniba/LLaMAntino-2-7b-hf-ITA"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

prompt = "Scrivi qui un possibile prompt"

input_ids = tokenizer(prompt, return_tensors="pt").input_ids
outputs = model.generate(input_ids=input_ids)

print(tokenizer.batch_decode(outputs.detach().cpu().numpy()[:, input_ids.shape[1]:], skip_special_tokens=True)[0])
```

If you are facing issues when loading the model, you can try to load it quantized:

```python
model = AutoModelForCausalLM.from_pretrained(model_id, load_in_8bit=True)
```

*Note*: The model loading strategy above requires the [*bitsandbytes*](https://pypi.org/project/bitsandbytes/) and [*accelerate*](https://pypi.org/project/accelerate/) libraries

## Citation

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

If you use this model in your research, please cite the following:

```bibtex
@misc{basile2023llamantino,
      title={LLaMAntino: LLaMA 2 Models for Effective Text Generation in Italian Language}, 
      author={Pierpaolo Basile and Elio Musacchio and Marco Polignano and Lucia Siciliani and Giuseppe Fiameni and Giovanni Semeraro},
      year={2023},
      eprint={2312.09993},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```

*Notice:* Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved. [*License*](https://ai.meta.com/llama/license/)