File size: 1,763 Bytes
3349d47
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
---
license: llama2
pipeline_tag: text-generation
Tags:
- cortex.cpp
- multimodal
- vicuna
- vision-language
---

## Overview

**LLaVA** (Large Language and Vision Assistant) is an open-source chatbot trained to handle multimodal instruction-following tasks. It is a fine-tuned **Vicuna-7B** model, designed to process both **text and image** inputs. This auto-regressive language model leverages the **transformer architecture** to improve interactions in vision-language tasks, making it useful for research in **computer vision, natural language processing, machine learning, and artificial intelligence**.

LLaVA-v1.6-Vicuna-7B is the latest iteration, trained in **December 2023**, and optimized for improved instruction-following performance in multimodal settings.

## Variants

| No | Variant | Cortex CLI command |
| --- | --- | --- |
| 1 | [llava-v1.6-vicuna-7b-f16](https://huggingface.co/cortexso/llava-v1.6/tree/gguf-f16) | `cortex run llava-v1.6:gguf-f16` |
| 2 | [llava-v1.6-vicuna-7b-q4_km](https://huggingface.co/cortexso/llava-v1.6/tree/gguf-q4-km) | `cortex run llava-v1.6:gguf-q4-km` |

## Use it with Jan (UI)

1. Install **Jan** using [Quickstart](https://jan.ai/docs/quickstart)
2. Use in Jan model Hub:
    ```bash
    cortexso/llava-v1.6
    ```

## Use it with Cortex (CLI)

1. Install **Cortex** using [Quickstart](https://cortex.jan.ai/docs/quickstart)
2. Run the model with command:
    ```bash
    cortex run llava-v1.6
    ```

## Credits
- **Author:** LLaVA Research Team
- **Converter:** [Homebrew](https://www.homebrew.ltd/)
- **Original License:** [LLAMA 2 Community License](https://github.com/facebookresearch/llama/blob/main/LICENSE)
- **Papers:** [LLaVA-v1.6: Enhancing Large Multimodal Models](https://llava-vl.github.io/)