Image-Text-to-Text
Transformers
Safetensors
qwen2_5_vl
autonomous-driving
vision-language-action
trajectory-prediction
navsim
vllm
conversational
text-generation-inference
Instructions to use AppliedIntuitionResearch/nord-base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AppliedIntuitionResearch/nord-base with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="AppliedIntuitionResearch/nord-base") 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)# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("AppliedIntuitionResearch/nord-base") model = AutoModelForMultimodalLM.from_pretrained("AppliedIntuitionResearch/nord-base", 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=256) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use AppliedIntuitionResearch/nord-base with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "AppliedIntuitionResearch/nord-base" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "AppliedIntuitionResearch/nord-base", "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/AppliedIntuitionResearch/nord-base
- SGLang
How to use AppliedIntuitionResearch/nord-base 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 "AppliedIntuitionResearch/nord-base" \ --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": "AppliedIntuitionResearch/nord-base", "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 "AppliedIntuitionResearch/nord-base" \ --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": "AppliedIntuitionResearch/nord-base", "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 AppliedIntuitionResearch/nord-base with Docker Model Runner:
docker model run hf.co/AppliedIntuitionResearch/nord-base
Update title, GitHub org, and description wording
Browse files
README.md
CHANGED
|
@@ -11,16 +11,17 @@ tags:
|
|
| 11 |
- vllm
|
| 12 |
---
|
| 13 |
|
| 14 |
-
# NoRD-Base:
|
| 15 |
|
| 16 |
-
**CVPR 2026** | [arXiv](https://arxiv.org/abs/2602.21172) | [Project Page](https://nord-vla-ai.github.io/) | [GitHub](https://github.com/
|
| 17 |
|
| 18 |
*Ishaan Rawal · Shubh Gupta · Yihan Hu · Wei Zhan*
|
| 19 |
|
| 20 |
This is the **SFT-only ablation baseline** from the paper: [Qwen2.5-VL-3B](https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct)
|
| 21 |
-
supervised fine-tuned to directly predict driving trajectories as discrete tokens
|
| 22 |
-
chain-of-thought reasoning, 3× fewer tokens than reasoning-based VLA baselines. This
|
| 23 |
-
is also the initialization used for the Dr. GRPO stage that produces the paper's
|
|
|
|
| 24 |
|
| 25 |
| Model | Training | NAVSIM PDMS |
|
| 26 |
|---|---|---|
|
|
@@ -29,7 +30,7 @@ is also the initialization used for the Dr. GRPO stage that produces the paper's
|
|
| 29 |
|
| 30 |
## Usage
|
| 31 |
|
| 32 |
-
Install the [`nord`](https://github.com/
|
| 33 |
|
| 34 |
```bash
|
| 35 |
pip install "nord[serve]"
|
|
@@ -51,7 +52,7 @@ output = agent.predict(nord.NordInput(
|
|
| 51 |
print(output.trajectory.shape) # (40, 3) — x, y, heading at 10 Hz, 4 seconds
|
| 52 |
```
|
| 53 |
|
| 54 |
-
Full inference and NAVSIM evaluation instructions: [github.com/
|
| 55 |
|
| 56 |
This repo bundles the K-Disc trajectory tokenizer vocab (`vocab.pkl`, 2048 clusters) used to
|
| 57 |
decode the model's output tokens into `(x, y, heading)` trajectories.
|
|
@@ -71,4 +72,4 @@ decode the model's output tokens into `(x, y, heading)` trajectories.
|
|
| 71 |
|
| 72 |
This checkpoint is released under [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/)
|
| 73 |
(non-commercial). The `nord` inference client code is separately licensed under Apache 2.0 —
|
| 74 |
-
see [github.com/
|
|
|
|
| 11 |
- vllm
|
| 12 |
---
|
| 13 |
|
| 14 |
+
# NoRD-Base: A Data-Efficient Vision-Language-Action Model that Drives without Reasoning (SFT only)
|
| 15 |
|
| 16 |
+
**CVPR 2026** | [arXiv](https://arxiv.org/abs/2602.21172) | [Project Page](https://nord-vla-ai.github.io/) | [GitHub](https://github.com/Applied-Intuition-Open-Source/nord)
|
| 17 |
|
| 18 |
*Ishaan Rawal · Shubh Gupta · Yihan Hu · Wei Zhan*
|
| 19 |
|
| 20 |
This is the **SFT-only ablation baseline** from the paper: [Qwen2.5-VL-3B](https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct)
|
| 21 |
+
supervised fine-tuned to directly predict driving trajectories as discrete tokens without
|
| 22 |
+
chain-of-thought reasoning, using 3× fewer tokens than reasoning-based VLA baselines. This
|
| 23 |
+
checkpoint is also the initialization used for the Dr. GRPO stage that produces the paper's
|
| 24 |
+
main result.
|
| 25 |
|
| 26 |
| Model | Training | NAVSIM PDMS |
|
| 27 |
|---|---|---|
|
|
|
|
| 30 |
|
| 31 |
## Usage
|
| 32 |
|
| 33 |
+
Install the [`nord`](https://github.com/Applied-Intuition-Open-Source/nord) client and serve with vLLM:
|
| 34 |
|
| 35 |
```bash
|
| 36 |
pip install "nord[serve]"
|
|
|
|
| 52 |
print(output.trajectory.shape) # (40, 3) — x, y, heading at 10 Hz, 4 seconds
|
| 53 |
```
|
| 54 |
|
| 55 |
+
Full inference and NAVSIM evaluation instructions: [github.com/Applied-Intuition-Open-Source/nord](https://github.com/Applied-Intuition-Open-Source/nord).
|
| 56 |
|
| 57 |
This repo bundles the K-Disc trajectory tokenizer vocab (`vocab.pkl`, 2048 clusters) used to
|
| 58 |
decode the model's output tokens into `(x, y, heading)` trajectories.
|
|
|
|
| 72 |
|
| 73 |
This checkpoint is released under [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/)
|
| 74 |
(non-commercial). The `nord` inference client code is separately licensed under Apache 2.0 —
|
| 75 |
+
see [github.com/Applied-Intuition-Open-Source/nord](https://github.com/Applied-Intuition-Open-Source/nord).
|