Instructions to use axteam/PaddleOCR-VL-FP8-dynamic with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use axteam/PaddleOCR-VL-FP8-dynamic with Transformers:

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

pipe = pipeline("image-text-to-text", model="axteam/PaddleOCR-VL-FP8-dynamic", 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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("axteam/PaddleOCR-VL-FP8-dynamic", trust_remote_code=True)
model = AutoModelForMultimodalLM.from_pretrained("axteam/PaddleOCR-VL-FP8-dynamic", trust_remote_code=True, 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 axteam/PaddleOCR-VL-FP8-dynamic with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "axteam/PaddleOCR-VL-FP8-dynamic"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "axteam/PaddleOCR-VL-FP8-dynamic",
		"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/axteam/PaddleOCR-VL-FP8-dynamic

SGLang

How to use axteam/PaddleOCR-VL-FP8-dynamic 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 "axteam/PaddleOCR-VL-FP8-dynamic" \
    --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": "axteam/PaddleOCR-VL-FP8-dynamic",
		"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 "axteam/PaddleOCR-VL-FP8-dynamic" \
        --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": "axteam/PaddleOCR-VL-FP8-dynamic",
		"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 axteam/PaddleOCR-VL-FP8-dynamic with Docker Model Runner:
```
docker model run hf.co/axteam/PaddleOCR-VL-FP8-dynamic
```

PaddleOCR-VL · FP8 (dynamic)

8-bit quantization of PaddlePaddle/PaddleOCR-VL for vLLM serving. Quantized by axteam.


Base model	`PaddlePaddle/PaddleOCR-VL`
Method	FP8 (dynamic)
Format	compressed-tensors `FP8_DYNAMIC`
Weights / activations	W8A8 — FP8 E4M3 weights (per-channel), dynamic per-token FP8 activations
Architecture	Dense (ERNIE-4.5 text backbone) · SigLIP-style vision encoder + projector · document/OCR VLM
Size on disk	~1.5 GB
Calibration	None (data-free)
License	apache-2.0

Quantization details

Scheme — W8A8 — FP8 E4M3 weights (per-channel), dynamic per-token FP8 activations, stored as compressed-tensors FP8_DYNAMIC.
Kept in BF16 — the vision tower (visual.*), the vision-text projector (mlp_AR.*), token embeddings, the LM head, and all RMSNorms (a small fraction of parameters). Only the language-model backbone's attention & MLP projections are quantized, so the OCR/vision path is untouched.
Post-training quantization (no retraining). Tokenizer / chat_template / processor identical to the original.

Evaluation (measured)

Fidelity to the BF16 original on an internal mixed EN/KO set (100 prompts, text next-token scoring).

Metric	This model	BF16 original
Perplexity (lower = closer)	1.1050	1.1059 (-0.1%)
Top-1 next-token agreement	99.6%	100.0%

Usage (vLLM)

vllm serve axteam/PaddleOCR-VL-FP8-dynamic --trust-remote-code

License

Quantized derivative of PaddlePaddle/PaddleOCR-VL (apache-2.0). Original weights & processor belong to the upstream authors.

한국어 요약

PaddlePaddle/PaddleOCR-VL 를 FP8 (dynamic) 로 양자화한 서빙용(8-bit) OCR 모델입니다. 작성: axteam.

방식 — FP8 E4M3, 가중치 채널별 스케일(정적·보정불필요), 활성화 동적 FP8. 형식 compressed-tensors FP8_DYNAMIC.
BF16 유지 — 비전 타워(visual.*)·비전-텍스트 프로젝터(mlp_AR.*)·토큰 임베딩·LM 헤드·모든 norm. 언어모델 백본의 attention·MLP projection 만 양자화(비전/OCR 경로 그대로).
PTQ(재학습 없음), 토크나이저·chat_template·프로세서 원본 동일.
품질(실측) — BF16 대비 perplexity 1.1050(원본 1.1059, -0.1%), top-1 99.6%(원본 100.0%). 내부 영/한 100문항.
사용 — vllm serve axteam/PaddleOCR-VL-FP8-dynamic --trust-remote-code. 라이선스 apache-2.0.