Instructions to use dots-studio/dots3-note-prev with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dots-studio/dots3-note-prev with Transformers:

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

pipe = pipeline("image-text-to-text", model="dots-studio/dots3-note-prev")
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 AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("dots-studio/dots3-note-prev", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use dots-studio/dots3-note-prev with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dots-studio/dots3-note-prev"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dots-studio/dots3-note-prev",
		"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/dots-studio/dots3-note-prev

SGLang

How to use dots-studio/dots3-note-prev 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 "dots-studio/dots3-note-prev" \
    --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": "dots-studio/dots3-note-prev",
		"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 "dots-studio/dots3-note-prev" \
        --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": "dots-studio/dots3-note-prev",
		"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 dots-studio/dots3-note-prev with Docker Model Runner:
```
docker model run hf.co/dots-studio/dots3-note-prev
```

Add community evaluation results

by SaylorTwift HF Staff - opened about 8 hours ago

base: refs/heads/main

←

from: refs/pr/5

Discussion Files changed

about 8 hours ago

No description provided.

Add community evaluation resultse8d2aa51

SaylorTwift

about 8 hours ago

Hugging Face CLI

Summary

This PR adds evaluation results extracted from the model card's benchmark charts and appendix tables (images, not text tables) to .eval_results/, following the Hugging Face Hub evaluation-results specification.

Benchmarks Added

SWE-bench Verified — 78.4
SWE-bench Multilingual — 75.7
SWE-bench Pro — 61.0
Claw-Eval — 73.4
WildClawBench — 61.7
MMMU-Pro — 79.1
Humanity's Last Exam (w/ tool) — 52.6
APEX-Agents — 30.8
Video-MME-v2 — 39.3
SkillsBench v1.1 — 52.8

Mapping Notes

Claw-Eval: the card reports one aggregate Pass³ score (N=3 trials) without a general/multimodal/multi_turn split; mapped to the general task as the most likely default.
SWE-bench suite (Verified, Multilingual, Pro): all three were run with the model's own "live-swe-agent" scaffold, temperature=1.0, top_p=0.95, max_new_tokens=32K, 384K context — noted once here rather than on each entry since it's constant across the three.
SkillsBench: 52.8 is the with-skills condition, not the with/without-skills lift (SkillsBench's other headline metric).

Benchmarks Skipped (Not Registered on Hub)

These appeared on the model card but have no registered eval.yaml on the Hub, so they were left out of the YAML:

IMOAnswerBench (90.9), Codeforces (3056 Elo), LiveCodeBench v6 (91.5), ARC-AGI-2 (81.4), ARC-AGI-3 arcagi3 harness (6.9), ARC-AGI-3 general harness (32.1), IFBench (80.4), IFEval (93.9), VibeLifeBench (28.1), NL2repo (49.8), Toolathlon-Verified (55.6), BrowseComp w/ CM (83.3), BrowseComp-zh (75.8), LiveBrowseComp (46.9), WideSearch (78.9), DeepSearchQA (92.1), VibeSearchBench (25.7), WorldVQA (42.7), SimpleVQA (72.5), MathVision (87.7), HiPho (74.8), ZeroBench@5 (19.0), CharxivReasoning (83.1), GDP.pdf mean-criteria (60.7), PerceptionBench (53.4), BabyVision (50.0), MMVU (79.9), VideoMMMU (86.8), LongVideoBench (76.8), VideoZeroBench (17.4), VoiceBench (92.4), AudioMC (46.9), MMAU-PRO (69.8).

These can be added once the benchmark authors register an eval.yaml on the Hub.

Also Considered

Terminal-Bench (card value 75.1, registered as harborframework/terminal-bench-2.0 / terminalbench_2) was left out of this PR.

Source

Model card: https://huggingface.co/dots-studio/dots3-note-prev
Benchmark images: assets/bench_en1.png, assets/bench_en2.png, assets/benchmark_appendix_en_reasoning.png, assets/benchmark_appendix_en_multimodal.png
Paper: not yet published ("Full Report (coming soon)" per the model card)

Verification

These results were read directly off the model card's published benchmark charts and appendix tables (visual read, not text extraction). No verified token is attached, as these were not run via HF Jobs with inspect-ai.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

Ready to merge

This branch is ready to get merged automatically.

· Sign up or log in to comment