GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
SujieHu
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
SujieHu
Model Tree
Qwen/Qwen3-VL-8B-Instruct
Input Modalities
Output Modalities
Supported Functionality
TextDART returns one quadrilateral and transcription for each visible text region:
json
{ "regions": [ { "quad": [x1, y1, x2, y2, x3, y3, x4, y4], "recognized_text": "recognized text" } ]}Coordinates are normalized to a 0–1000 grid. The model uses the following structural
anomaly markers:
<#>: one malformed and unreadable Chinese character or English letter<###>: one malformed and unreadable English word<#####>: an entirely unreadable block of secondary or fine-print textThe official repository contains the exact system prompt, an OpenAI-compatible vLLM deployment script, response parsing, and visualization:
bash
git clone https://github.com/WeChatCV/TextDART-SCANNER.gitcd TextDART-SCANNER/TextDART/train conda create -n TextDART python=3.11 -yconda activate TextDARTbash install_all.sh # Each listed GPU hosts one full model replica.bash deploy_server.sh SujieHu/TextDART-8B-Qwen3VL 1925 0In another shell:
bash
conda activate TextDARTcd TextDART-SCANNER/TextDART/trainpython infer_demo.py --server-url http://127.0.0.1:1925The demo processes the included example images and writes parsed JSON plus annotated images. See the TextDART deployment guide for custom images, server options, and training instructions. See the main repository for using TextDART as the reward model in SCANNER.
TextDART is intended for:
If you find this model useful, please cite:
bibtex
@article{hu2026textdartscanner, title={Blame Errant Words, Not the Canvas: Defect-Aware Spatial Credit Assignment for Visual Text Generation}, author={Hu, Sujie and Zhang, Huaisong and Yang, Pujing and Wang, Qian and Li, Xiaomin and Leng, Xinhang and Li, Chen and Lyu, Jing and Li, Xiu and Pan, Ling}, journal={arXiv preprint arXiv:2610.20794}, year={2026}}{ "regions": [ { "quad": [x1, y1, x2, y2, x3, y3, x4, y4], "recognized_text": "recognized text" } ]}git clone https://github.com/WeChatCV/TextDART-SCANNER.gitcd TextDART-SCANNER/TextDART/train conda create -n TextDART python=3.11 -yconda activate TextDARTbash install_all.sh # Each listed GPU hosts one full model replica.bash deploy_server.sh SujieHu/TextDART-8B-Qwen3VL 1925 0conda activate TextDARTcd TextDART-SCANNER/TextDART/trainpython infer_demo.py --server-url http://127.0.0.1:1925@article{hu2026textdartscanner, title={Blame Errant Words, Not the Canvas: Defect-Aware Spatial Credit Assignment for Visual Text Generation}, author={Hu, Sujie and Zhang, Huaisong and Yang, Pujing and Wang, Qian and Li, Xiaomin and Leng, Xinhang and Li, Chen and Lyu, Jing and Li, Xiu and Pan, Ling}, journal={arXiv preprint arXiv:2610.20794}, year={2026}}