Image-Text-to-Text
Transformers
Safetensors
multilingual
deepseek_vl_v2
feature-extraction
deepseek
vision-language
ocr
custom_code
Eval Results
Instructions to use deepseek-ai/DeepSeek-OCR with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use deepseek-ai/DeepSeek-OCR with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="deepseek-ai/DeepSeek-OCR", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("deepseek-ai/DeepSeek-OCR", trust_remote_code=True, device_map="auto") - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use deepseek-ai/DeepSeek-OCR with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "deepseek-ai/DeepSeek-OCR" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "deepseek-ai/DeepSeek-OCR", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/deepseek-ai/DeepSeek-OCR
- SGLang
How to use deepseek-ai/DeepSeek-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 "deepseek-ai/DeepSeek-OCR" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "deepseek-ai/DeepSeek-OCR", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'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 "deepseek-ai/DeepSeek-OCR" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "deepseek-ai/DeepSeek-OCR", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use deepseek-ai/DeepSeek-OCR with Docker Model Runner:
docker model run hf.co/deepseek-ai/DeepSeek-OCR
Clean up Real5-OmniDocBench evaluation metadata
Browse filesRemove result footnotes and point source metadata to the public benchmark repository. Evaluation scores are unchanged.
.eval_results/real5_omnidocbench.yaml
CHANGED
|
@@ -3,52 +3,46 @@
|
|
| 3 |
task_id: overall
|
| 4 |
value: 73.99
|
| 5 |
source:
|
| 6 |
-
url: https://
|
| 7 |
-
name: Real5-OmniDocBench
|
| 8 |
user: PaddlePaddle
|
| 9 |
-
notes: Deepseek-OCR; official Real5-OmniDocBench result
|
| 10 |
- dataset:
|
| 11 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 12 |
task_id: scanning
|
| 13 |
value: 86.17
|
| 14 |
source:
|
| 15 |
-
url: https://
|
| 16 |
-
name: Real5-OmniDocBench
|
| 17 |
user: PaddlePaddle
|
| 18 |
-
notes: Deepseek-OCR; official Real5-OmniDocBench result
|
| 19 |
- dataset:
|
| 20 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 21 |
task_id: warping
|
| 22 |
value: 67.2
|
| 23 |
source:
|
| 24 |
-
url: https://
|
| 25 |
-
name: Real5-OmniDocBench
|
| 26 |
user: PaddlePaddle
|
| 27 |
-
notes: Deepseek-OCR; official Real5-OmniDocBench result
|
| 28 |
- dataset:
|
| 29 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 30 |
task_id: screen_photography
|
| 31 |
value: 75.31
|
| 32 |
source:
|
| 33 |
-
url: https://
|
| 34 |
-
name: Real5-OmniDocBench
|
| 35 |
user: PaddlePaddle
|
| 36 |
-
notes: Deepseek-OCR; official Real5-OmniDocBench result
|
| 37 |
- dataset:
|
| 38 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 39 |
task_id: illumination
|
| 40 |
value: 78.1
|
| 41 |
source:
|
| 42 |
-
url: https://
|
| 43 |
-
name: Real5-OmniDocBench
|
| 44 |
user: PaddlePaddle
|
| 45 |
-
notes: Deepseek-OCR; official Real5-OmniDocBench result
|
| 46 |
- dataset:
|
| 47 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 48 |
task_id: skew
|
| 49 |
value: 63.01
|
| 50 |
source:
|
| 51 |
-
url: https://
|
| 52 |
-
name: Real5-OmniDocBench
|
| 53 |
user: PaddlePaddle
|
| 54 |
-
notes: Deepseek-OCR; official Real5-OmniDocBench result
|
|
|
|
| 3 |
task_id: overall
|
| 4 |
value: 73.99
|
| 5 |
source:
|
| 6 |
+
url: https://huggingface.co/datasets/PaddlePaddle/Real5-OmniDocBench
|
| 7 |
+
name: Real5-OmniDocBench Leaderboard
|
| 8 |
user: PaddlePaddle
|
|
|
|
| 9 |
- dataset:
|
| 10 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 11 |
task_id: scanning
|
| 12 |
value: 86.17
|
| 13 |
source:
|
| 14 |
+
url: https://huggingface.co/datasets/PaddlePaddle/Real5-OmniDocBench
|
| 15 |
+
name: Real5-OmniDocBench Leaderboard
|
| 16 |
user: PaddlePaddle
|
|
|
|
| 17 |
- dataset:
|
| 18 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 19 |
task_id: warping
|
| 20 |
value: 67.2
|
| 21 |
source:
|
| 22 |
+
url: https://huggingface.co/datasets/PaddlePaddle/Real5-OmniDocBench
|
| 23 |
+
name: Real5-OmniDocBench Leaderboard
|
| 24 |
user: PaddlePaddle
|
|
|
|
| 25 |
- dataset:
|
| 26 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 27 |
task_id: screen_photography
|
| 28 |
value: 75.31
|
| 29 |
source:
|
| 30 |
+
url: https://huggingface.co/datasets/PaddlePaddle/Real5-OmniDocBench
|
| 31 |
+
name: Real5-OmniDocBench Leaderboard
|
| 32 |
user: PaddlePaddle
|
|
|
|
| 33 |
- dataset:
|
| 34 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 35 |
task_id: illumination
|
| 36 |
value: 78.1
|
| 37 |
source:
|
| 38 |
+
url: https://huggingface.co/datasets/PaddlePaddle/Real5-OmniDocBench
|
| 39 |
+
name: Real5-OmniDocBench Leaderboard
|
| 40 |
user: PaddlePaddle
|
|
|
|
| 41 |
- dataset:
|
| 42 |
id: PaddlePaddle/Real5-OmniDocBench
|
| 43 |
task_id: skew
|
| 44 |
value: 63.01
|
| 45 |
source:
|
| 46 |
+
url: https://huggingface.co/datasets/PaddlePaddle/Real5-OmniDocBench
|
| 47 |
+
name: Real5-OmniDocBench Leaderboard
|
| 48 |
user: PaddlePaddle
|
|
|