Instructions to use ibm-granite/granite-docling-258M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ibm-granite/granite-docling-258M with Transformers:

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

pipe = pipeline("image-text-to-text", model="ibm-granite/granite-docling-258M")
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("ibm-granite/granite-docling-258M")
model = AutoModelForMultimodalLM.from_pretrained("ibm-granite/granite-docling-258M", 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 ibm-granite/granite-docling-258M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ibm-granite/granite-docling-258M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ibm-granite/granite-docling-258M",
		"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/ibm-granite/granite-docling-258M

SGLang

How to use ibm-granite/granite-docling-258M 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 "ibm-granite/granite-docling-258M" \
    --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": "ibm-granite/granite-docling-258M",
		"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 "ibm-granite/granite-docling-258M" \
        --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": "ibm-granite/granite-docling-258M",
		"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 ibm-granite/granite-docling-258M with Docker Model Runner:
```
docker model run hf.co/ibm-granite/granite-docling-258M
```

no simple compatible

#44

by kalle07 - opened Jan 2

Discussion

kalle07

Jan 2

•

edited Jan 2

i get all models to run, qwen, qwen_vl, next_ocr, embedder models ...all

but yours sucks

transformers 4.57.3
i must read out to find out
Idefics3Processor, Idefics3ForConditionalGeneration

AND
vllm cant be run simple with windows
why all not compatible?

this dosnt work output is emptym, so no ocr

from transformers import Idefics3Processor, Idefics3ForConditionalGeneration
from PIL import Image
import torch

MODEL_PATH = r"f:\LLM-models\granite-docling"
IMAGE_PATH = r"f:\LLM-models\granite-docling\Unbenannt01.PNG"

processor = Idefics3Processor.from_pretrained(MODEL_PATH)
model = Idefics3ForConditionalGeneration.from_pretrained(MODEL_PATH)

device = "cuda" if torch.cuda.is_available() else "cpu"
model.to(device)

# Load image
image = Image.open(IMAGE_PATH).convert("RGB")

# Prompt must match the model’s expected instruction
prompt = ["OCR this document image: <image>"]

# Properly move all tensors to device
inputs = processor(text=prompt, images=[image], return_tensors="pt")
inputs = {k: v.to(device) for k, v in inputs.items()}

# Generate output
with torch.no_grad():
    outputs = model.generate(**inputs, max_new_tokens=1024, do_sample=False)

# Decode
text = processor.batch_decode(outputs, skip_special_tokens=True)[0]
print(text)

%Run -c $EDITOR_CONTENT
OCR this document image:

gabegoodhart

IBM Granite org Jan 5

Hi @kalle07 , it looks like you are using the prompt text directly rather than using apply_chat_template as shown here (see the transformers section): https://huggingface.co/ibm-granite/granite-docling-258M#getting-started.

This model is specifically designed to be used with the docling library, so it expects very specific prompt formatting which the library handles directly. When using it directly, the chat template must be used to get the expected input format correct (including things like trailing newline characters and special delimiter tokens).

vpakarinen

Mar 6

•

edited Mar 6

You are the one who sucks here, you don't even know how to use it properly.

There is nothing wrong with docling.

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

Tap or paste here to upload images

· Sign up or log in to comment