File size: 1,315 Bytes
2bbca78 1ec9ef9 5cccb25 2bbca78 5cccb25 2bbca78 ae19cec 2bbca78 5cccb25 2bbca78 42963ee 2bbca78 e06160b 5cccb25 2bbca78 5cccb25 2bbca78 42963ee 2bbca78 |
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 |
---
license: apache-2.0
inference: false
base_model: Qwen/Qwen2-VL-2B-Instruct
base_model_relation: quantized
tags: [green, llmware-vision, p2, ov, emerald]
---
# qwen2-vl-2b-instruct-ov
**qwen2-vl-2b-instruct-ov** is an OpenVino int4 quantized version of [Qwen2-VL-2B-Instruct](https://www.huggingface.co/Qwen/Qwen2-VL-2B-Instruct), providing a very fast, very small inference implementation, optimized for AI PCs using Intel GPU, CPU and NPU.
This is a vision-to-text model from the Qwen2 release series from Qwen (summer 2024), and is a very high-quality innovative small model that accepts multi-modal inputs (image/video, text).
### Model Description
- **Developed by:** Qwen
- **Quantized by:** llmware
- **Model type:** qwen2
- **Parameters:** 2.2 billion
- **Model Parent:** Qwen/Qwen2-VL-2B-Instruct
- **Language(s) (NLP):** English
- **License:** Apache 2.0
- **Uses:** Multimodal LLM
- **Quantization:** int4
For an open source inference implementation, please see this [Intel OpenVino notebook](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/qwen2-vl)
## Model Card Contact
[llmware on github](https://www.github.com/llmware-ai/llmware)
[llmware on hf](https://www.huggingface.co/llmware)
[llmware website](https://www.llmware.ai)
|