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

Libraries

How to use dots-studio/dots.ocr with Transformers:

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

pipe = pipeline("image-text-to-text", model="dots-studio/dots.ocr", 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 AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("dots-studio/dots.ocr", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use dots-studio/dots.ocr with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dots-studio/dots.ocr"
# 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/dots.ocr",
		"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/dots.ocr

SGLang

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

Guard `cache_position` against `None` in `prepare_inputs_for_generation`

#52

by wittjeff - opened 2 days ago

base: refs/heads/main

←

from: refs/pr/52

Discussion Files changed

-2

wittjeff

2 days ago

Guard cache_position against None in prepare_inputs_for_generation

prepare_inputs_for_generation subscripts cache_position unconditionally:

if cache_position[0] == 0:
    model_inputs["pixel_values"] = pixel_values

Recent transformers releases pass cache_position=None on some generation paths, which crashes with TypeError: 'NoneType' object is not subscriptable at this line. The failure is content-dependent and version-dependent, which makes it confusing to hit in practice:

transformers 5.14: fails on a subset of inputs — in our document-OCR benchmarking, 4 of 25 OmniDocBench pages crashed while the rest processed normally.
transformers 5.12: fails immediately on the first generation call.

Full traceback (5.12, but the failing frame is identical on 5.14):

File ".../transformers/generation/utils.py", line 3824, in _prefill
  model_inputs = self.prepare_inputs_for_generation(
File ".../modeling_dots_ocr.py", line 128, in prepare_inputs_for_generation
  if cache_position[0] == 0:
TypeError: 'NoneType' object is not subscriptable

This PR adds the standard guard used by in-tree transformers models:

if cache_position is not None and cache_position[0] == 0:

Semantics are preserved: when cache_position is None, pixel_values is simply not re-attached for that step, matching the behavior of the equivalent guard in upstream transformers VLM implementations. Verified on the previously-crashing pages — they now generate and parse normally.

Guard `cache_position` against `None` in `prepare_inputs_for_generation`ac10f9c0

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