Instructions to use Phu-Hien/LightOnOCR-2-ft-04 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Phu-Hien/LightOnOCR-2-ft-04 with Transformers:

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

pipe = pipeline("image-text-to-text", model="Phu-Hien/LightOnOCR-2-ft-04")
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, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("Phu-Hien/LightOnOCR-2-ft-04")
model = AutoModelForImageTextToText.from_pretrained("Phu-Hien/LightOnOCR-2-ft-04")
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

vLLM

How to use Phu-Hien/LightOnOCR-2-ft-04 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Phu-Hien/LightOnOCR-2-ft-04"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Phu-Hien/LightOnOCR-2-ft-04",
		"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/Phu-Hien/LightOnOCR-2-ft-04

SGLang

How to use Phu-Hien/LightOnOCR-2-ft-04 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 "Phu-Hien/LightOnOCR-2-ft-04" \
    --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": "Phu-Hien/LightOnOCR-2-ft-04",
		"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 "Phu-Hien/LightOnOCR-2-ft-04" \
        --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": "Phu-Hien/LightOnOCR-2-ft-04",
		"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 Phu-Hien/LightOnOCR-2-ft-04 with Docker Model Runner:
```
docker model run hf.co/Phu-Hien/LightOnOCR-2-ft-04
```

LightOnOCR-2-ft-04

This model is a fine-tuned version of lightonai/LightOnOCR-2-1B-base on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.1808

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 6
eval_batch_size: 6
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 24
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 10
num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss
0.1814	0.3717	25	0.2257
0.1608	0.7435	50	0.2082
0.1349	1.1041	75	0.1988
0.1506	1.4758	100	0.1922
0.1205	1.8476	125	0.1882
0.1170	2.2082	150	0.1858
0.1150	2.5799	175	0.1842
0.1199	2.9517	200	0.1829
0.1110	3.3123	225	0.1819
0.1231	3.6840	250	0.1813
0.1158	4.0446	275	0.1809
0.1131	4.4164	300	0.1809
0.1287	4.7881	325	0.1808

Framework versions

Transformers 5.0.0
Pytorch 2.11.0+cu128
Datasets 4.8.5
Tokenizers 0.22.2

Downloads last month: 21

Safetensors

Model size

1B params

Tensor type

BF16

Model tree for Phu-Hien/LightOnOCR-2-ft-04

Base model

lightonai/LightOnOCR-2-1B-base

Finetuned

(17)

this model