Instructions to use wanglne/EVA-InternVL3.5-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use wanglne/EVA-InternVL3.5-8B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="wanglne/EVA-InternVL3.5-8B", trust_remote_code=True)
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("wanglne/EVA-InternVL3.5-8B", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use wanglne/EVA-InternVL3.5-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "wanglne/EVA-InternVL3.5-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "wanglne/EVA-InternVL3.5-8B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/wanglne/EVA-InternVL3.5-8B

SGLang

How to use wanglne/EVA-InternVL3.5-8B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "wanglne/EVA-InternVL3.5-8B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "wanglne/EVA-InternVL3.5-8B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "wanglne/EVA-InternVL3.5-8B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "wanglne/EVA-InternVL3.5-8B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use wanglne/EVA-InternVL3.5-8B with Docker Model Runner:
```
docker model run hf.co/wanglne/EVA-InternVL3.5-8B
```

EVA-InternVL3.5-8B

File size: 2,176 Bytes

d11e030

---
license: apache-2.0
base_model:
- OpenGVLab/InternVL3_5-8B
pipeline_tag: image-text-to-text
library_name: transformers
tags:
- EVA
- internvl
- model-editing
- ai-safety
- jailbreak
- defense
- vision-language-model
---

# EVA-InternVL3.5-8B

This repository contains the final InternVL3.5-8B checkpoint produced by **EVA**.

EVA is an editing-based framework for improving the robustness and safety alignment of large language models and vision-language models against diverse jailbreak attacks while preserving their general capabilities on benign tasks.

## Model information

- **EVA checkpoint:** `wanglne/EVA-InternVL3.5-8B`
- **Base model:** [OpenGVLab/InternVL3_5-8B](https://huggingface.co/OpenGVLab/InternVL3_5-8B)

## Resources

- **Paper:** [EVA: Editing for Versatile Alignment against Jailbreaks](https://arxiv.org/abs/2605.14750)
- **IEEE TPAMI:** [10.1109/TPAMI.2026.3694189](https://doi.org/10.1109/TPAMI.2026.3694189)
- **Code:** [wanglne/EVA](https://github.com/wanglne/EVA)
- **EVA model collection:** [EVA models](https://huggingface.co/collections/wanglne/eva-editing-for-versatile-alignment-against-jailbreaks)

## Intended use

This checkpoint is released for research, reproduction, safety evaluation, and comparison with other alignment methods.

## License

This checkpoint is derived from InternVL3.5-8B. Its use and redistribution remain subject to the Apache License 2.0 and the applicable base-model terms.

The EVA source code is released separately under the MIT License.

## Citation

```bibtex
@ARTICLE{11523146,
  author={Wang, Yi and Qiu, Hongye and Xu, Yue and Yang, Sibei and Qin, Zhan and Huang, Minlie and Wang, Wenjie},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence}, 
  title={EVA: Editing for Versatile Alignment against Jailbreaks}, 
  year={2026},
  volume={},
  number={},
  pages={1-16},
  keywords={Automatic speech recognition;Modeling;Safety;Visualization;Large language models;Conferences;Optimization;Educational institutions;Light emitting diodes;Tuning;Safety Alignment;Jailbreak Attacks;Model Editing;Large Language Models;Vision Language Models},
  doi={10.1109/TPAMI.2026.3694189}}
```