Image-Text-to-Text
Transformers
Safetensors
English
Chinese
qwen3_5_moe
exomind
scientific-reasoning
scientific-research
agentic
tool-use
multimodal
vision-language
qwen3.5
conversational
Instructions to use AI4SGI/ExoMind with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AI4SGI/ExoMind with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="AI4SGI/ExoMind") 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 AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("AI4SGI/ExoMind") model = AutoModelForMultimodalLM.from_pretrained("AI4SGI/ExoMind", device_map="auto") 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?"} ] }, ] inputs = processor.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use AI4SGI/ExoMind with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "AI4SGI/ExoMind" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "AI4SGI/ExoMind", "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/AI4SGI/ExoMind
- SGLang
How to use AI4SGI/ExoMind 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 "AI4SGI/ExoMind" \ --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": "AI4SGI/ExoMind", "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 "AI4SGI/ExoMind" \ --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": "AI4SGI/ExoMind", "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 AI4SGI/ExoMind with Docker Model Runner:
docker model run hf.co/AI4SGI/ExoMind
| # ExoMind Research Content and Brand Terms | |
| Unless a specific item states otherwise, these terms apply to `Paper.pdf`, the | |
| paper's substantive content, the research and evaluation content in `README.md`, | |
| `CITATION.cff`, and `docs/` (including `docs/benchmark-data.js` and | |
| `docs/performance-data.js`), the ExoMind name and logo, and ExoMind visual brand | |
| assets. | |
| Copyright © 2026 the applicable rights holders. All rights reserved, subject to | |
| the permissions below. | |
| ## Linking and citation | |
| No separate permission from the ExoMind rights holders is required to link to | |
| the official [project page](https://ai4sgi.github.io/ExoMind/) or official | |
| [technical report](./Paper.pdf), provide the paper title and bibliographic | |
| citation, or discuss and cite ExoMind without reproducing reserved material | |
| beyond what applicable law permits. | |
| ## Limited noncommercial promotional permission | |
| Subject to the conditions below, the applicable rights holders grant permission | |
| to reproduce the following materials solely for noncommercial communication or | |
| promotion specifically about ExoMind: | |
| - the paper abstract or a faithful translation of it; | |
| - brief attributed excerpts from the report or official ExoMind announcements; | |
| and | |
| - an unmodified ExoMind logo or official ExoMind figure. | |
| These uses must: | |
| 1. clearly identify the source as **ExoMind Team, Shanghai Artificial | |
| Intelligence Laboratory**; | |
| 2. include a link to the official project page or technical report when the | |
| format permits; | |
| 3. preserve the original meaning and presentation, apart from proportionate | |
| resizing or layout-neutral cropping; and | |
| 4. not imply endorsement, affiliation, authorship, or sponsorship. | |
| The materials may not be sold, used in merchandise or paid advertising, or | |
| redistributed as a standalone asset or dataset collection. "Noncommercial" | |
| means not primarily intended for commercial advantage or monetary compensation. | |
| This permission applies only to rights controlled by the applicable rights | |
| holders. It does not grant rights in third-party names, logos, or materials that | |
| may appear in an ExoMind figure or webpage. | |
| ## Uses requiring prior written permission | |
| Without prior written permission from the applicable rights holders, you may | |
| not: | |
| - host, mirror, or redistribute the complete technical-report PDF instead of | |
| linking to the official copy; | |
| - publish or distribute a translation, adaptation, or other derivative version | |
| of the technical report; | |
| - reproduce substantial portions of the paper's text, tables, figures, or other | |
| content in another manuscript or submission beyond properly attributed uses | |
| permitted by applicable law; | |
| - use ExoMind research content or brand assets to market an unrelated project, | |
| imply endorsement, or for other commercial purposes. | |
| No permission in these terms authorizes a recipient to submit the technical | |
| report, a modified version, or substantially copied content as another work or | |
| under false or incomplete authorship; claim authorship of ExoMind; or remove its | |
| attribution. | |
| ## Independent research and applicable law | |
| These terms do not restrict linking, proper citation, independent | |
| implementation, independently written follow-on research, or any use permitted | |
| by applicable law, including fair use or fair dealing. They do not transfer | |
| authorship or waive academic-integrity obligations. Customary news reporting | |
| and quotation remain governed by applicable law. | |
| For permissions beyond the scope above, contact the ExoMind Team through the | |
| official project repository. | |