{"id":30033612,"url":"https://github.com/skyworkai/unipic","last_synced_at":"2025-08-06T22:47:13.289Z","repository":{"id":307211074,"uuid":"1027015746","full_name":"SkyworkAI/UniPic","owner":"SkyworkAI","description":"Unified Autoregressive Modeling for Visual Understanding and Generation","archived":false,"fork":false,"pushed_at":"2025-08-06T04:15:38.000Z","size":11482,"stargazers_count":357,"open_issues_count":9,"forks_count":16,"subscribers_count":6,"default_branch":"main","last_synced_at":"2025-08-06T04:16:29.661Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/SkyworkAI.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-07-27T06:01:44.000Z","updated_at":"2025-08-06T04:15:41.000Z","dependencies_parsed_at":"2025-08-06T04:10:21.205Z","dependency_job_id":null,"html_url":"https://github.com/SkyworkAI/UniPic","commit_stats":null,"previous_names":["skyworkai/unipic"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/SkyworkAI/UniPic","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FUniPic","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FUniPic/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FUniPic/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FUniPic/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/SkyworkAI","download_url":"https://codeload.github.com/SkyworkAI/UniPic/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FUniPic/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":269169608,"owners_count":24372060,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-08-06T02:00:09.910Z","response_time":99,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-08-06T22:47:02.724Z","updated_at":"2025-08-06T22:47:13.235Z","avatar_url":"https://github.com/SkyworkAI.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003c!-- markdownlint-disable first-line-h1 --\u003e\n\u003c!-- markdownlint-disable html --\u003e\n\u003c!-- markdownlint-disable no-duplicate-header --\u003e\n\n\u003cdiv align=\"center\"\u003e\n  \u003ch1\u003e\u003cstrong\u003eSkywork-UniPic\u003c/strong\u003e\u003c/h1\u003e\n\u003c/div\u003e\n\n\u003cfont size=7\u003e\u003cdiv align='center' \u003e  [[🤗 UniPic checkpoint](https://huggingface.co/Skywork/Skywork-UniPic-1.5B)] [[📖 Tech Report](https://arxiv.org/abs/2508.03320)] \u003c/font\u003e \n\nWelcome to the Skywork-UniPic repository! This repository contains the model weights and implementation of our unified 1.5B-parameter autoregressive model that natively integrates image understanding, text-to-image generation, and image editing capabilities within a single architecture.\n\n\n\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"teaser.png\" alt=\"Skywork UniPic Teaser\" width=\"90%\"\u003e\n\u003c/div\u003e\n\n## Evaluation\n\n\u003cp align=\"center\"\u003e\u003cstrong\u003ePerformance Overview\u003c/strong\u003e\u003c/p\u003e\n\n\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"./imgs/main_comparison.png\" alt=\"Main comparison\" width=\"800\"/\u003e\n\u003c/div\u003e\n\n\u003cp align=\"center\"\u003e\u003cstrong\u003eGenEval\u003c/strong\u003e\u003c/p\u003e\n\u003cdiv align=\"center\"\u003e\n\n\n| Model | Single | Two | Count | Color | Position | Attr | **Overall** |\n|:------|:------:|:---:|:-----:|:-----:|:--------:|:----:|:-----------:|\n| **Diffusion Models** |||||||||\n| SDv2.1          | 0.98 | 0.51 | 0.44 | 0.85 | 0.07 | 0.17 | 0.50 |\n| SDXL            | 0.98 | 0.74 | 0.39 | 0.85 | 0.15 | 0.23 | 0.55 |\n| IF‑XL           | 0.97 | 0.74 | 0.66 | 0.81 | 0.13 | 0.35 | 0.61 |\n| LUMINA‑Next     | 0.92 | 0.46 | 0.48 | 0.70 | 0.09 | 0.13 | 0.46 |\n| SD3‑medium      | 0.99 | 0.94 | 0.72 | 0.89 | 0.33 | 0.60 | 0.74 |\n| FLUX.1‑dev      | 0.99 | 0.81 | 0.79 | 0.74 | 0.20 | 0.47 | 0.67 |\n| NOVA            | 0.99 | 0.91 | 0.62 | 0.85 | 0.33 | 0.56 | 0.71 |\n| **Autoregressive Models** |||||||||\n| TokenFlow‑XL    | 0.95 | 0.60 | 0.41 | 0.81 | 0.16 | 0.24 | 0.55 |\n| Janus           | 0.97 | 0.68 | 0.30 | 0.84 | 0.46 | 0.42 | 0.61 |\n| Janus Pro       | 0.99 | 0.89 | 0.59 | 0.90 | 0.79 | 0.66 | 0.80 |\n| Emu3‑Gen        | 0.99 | 0.81 | 0.42 | 0.80 | 0.49 | 0.45 | 0.66 |\n| Show‑o          | 0.98 | 0.80 | 0.66 | 0.84 | 0.31 | 0.50 | 0.68 |\n| **Unified Models** |||||||||\n| OmniGen         | 0.98 | 0.84 | 0.66 | 0.74 | 0.40 | 0.43 | 0.68 |\n| OmniGen2        | 1.00 | 0.95 | 0.64 | 0.88 | 0.55 | 0.76 | 0.80 |\n| OmniGen2†       | 0.99 | 0.96 | 0.74 | 0.98 | 0.71 | 0.75 | 0.86 |\n| MetaQuery‑XL†   |   -  |   -  |   -  |   -  |   -  |   -  | 0.80 |\n| BLIP3‑o† 4B     |   -  |   -  |   -  |   -  |   -  |   -  | 0.81 |\n| BLIP3‑o† 8B     |   -  |   -  |   -  |   -  |   -  |   -  | 0.84 |\n| BAGEL           | 0.99 | 0.94 | 0.81 | 0.88 | 0.64 | 0.63 | 0.82 |\n| BAGEL†          | 0.98 | 0.95 | 0.84 | 0.95 | 0.78 | 0.77 | 0.88 |\n| UniWorld‑V1     | 0.99 | 0.93 | 0.79 | 0.89 | 0.49 | 0.70 | 0.80 |\n| UniWorld‑V1†    | 0.98 | 0.93 | 0.81 | 0.89 | 0.74 | 0.71 | 0.84 |\n| Ovis‑U1         | 0.98 | 0.98 | 0.90 | 0.92 | 0.79 | 0.75 | 0.89 |\n| **Proprietary Models** |||||||||\n| GPT‑4o          | 0.99 | 0.92 | 0.85 | 0.92 | 0.75 | 0.61 | 0.84 |\n| **Skywork UniPic** | **0.98** | **0.92** | **0.74** | **0.91** | **0.89** | **0.72** | **0.86** |\n\n\u003c/div\u003e\n\n\n\u003cp align=\"center\"\u003e\u003cstrong\u003eDPG‑Bench\u003c/strong\u003e\u003c/p\u003e\n\u003cdiv align=\"center\"\u003e\n\n| Model | Global | Entity | Attribute | Relation | Other | **Overall** |\n|:------|:------:|:------:|:---------:|:--------:|:-----:|:-----------:|\n| **Diffusion Models** |||||||\n| LUMINA‑Next      | 82.82 | 88.65 | 86.44 | 80.53 | 81.82 | 74.63 |\n| SDXL             | 83.27 | 82.43 | 80.91 | 86.76 | 80.41 | 74.65 |\n| PlayGroundv2.5   | 83.06 | 82.59 | 81.20 | 84.08 | 83.50 | 75.47 |\n| Hunyuan‑DiT      | 84.59 | 80.59 | 88.01 | 74.36 | 86.41 | 78.87 |\n| PixArt‑Σ         | 86.89 | 82.89 | 88.94 | 86.59 | 87.68 | 80.54 |\n| DALLE3           | 90.97 | 89.61 | 88.39 | 90.58 | 89.83 | 83.50 |\n| SD3‑medium       | 87.90 | 91.01 | 88.83 | 80.70 | 88.68 | 84.08 |\n| FLUX.1‑dev       | 82.10 | 89.50 | 88.70 | 91.10 | 89.40 | 84.00 |\n| **Autoregressive Models** |||||||\n| Show‑o           | 79.33 | 75.44 | 78.02 | 84.45 | 60.80 | 67.27 |\n| EMU3             | 85.21 | 86.68 | 86.84 | 90.22 | 83.15 | 80.60 |\n| TokenFlow‑XL     | 78.72 | 79.22 | 81.29 | 85.22 | 71.20 | 73.38 |\n| Janus            | 82.33 | 87.38 | 87.70 | 85.46 | 86.41 | 79.68 |\n| Janus Pro        | 86.90 | 88.90 | 89.40 | 89.32 | 89.48 | 84.19 |\n| BLIP3‑o 4B       |   -   |   -   |   -   |   -   |   -   | 79.36 |\n| BLIP3‑o 8B       |   -   |   -   |   -   |   -   |   -   | 81.60 |\n| **Unified Models** |||||||\n| OmniGen          | 87.90 | 88.97 | 88.47 | 87.95 | 83.56 | 81.16 |\n| OmniGen2         | 88.81 | 88.83 | 90.18 | 89.37 | 90.27 | 83.57 |\n| BAGEL            | 88.94 | 90.37 | 91.29 | 90.82 | 88.67 | 85.07 |\n| UniWorld‑V1      | 83.64 | 88.39 | 88.44 | 89.27 | 87.22 | 81.38 |\n| Ovis‑U1          | 82.37 | 90.08 | 88.68 | 93.35 | 85.20 | 83.72 |\n| **Skywork UniPic** | **89.65** | **87.78** | **90.84** | **91.89** | **91.95** | **85.50** |\n\n\u003c/div\u003e\n\nThe image editing feature within this UniPic 1.0 is an exploratory module at the forefront of research. It is not yet production-ready.\n\n\u003cp align=\"center\"\u003e\u003cstrong\u003eGEdit‑Bench‑EN\u003c/strong\u003e\u003c/p\u003e\n\u003cdiv align=\"center\"\u003e\n\n| Model | SC ↑ | PQ ↑ | **Overall ↑** |\n|:------|:----:|:----:|:--------------:|\n| **Proprietary Models** ||||\n| Gemini‑2.0‑flash | 6.73 | 6.61 | 6.32 |\n| GPT‑4o           | 7.85 | 7.62 | 7.53 |\n| **Specialized Editing Models** ||||\n| Instruct‑Pix2Pix | 3.58 | 5.49 | 3.68 |\n| MagicBrush       | 4.68 | 5.66 | 4.52 |\n| AnyEdit          | 3.18 | 5.82 | 3.21 |\n| ICEdit           | 5.11 | 6.85 | 4.84 |\n| Step1X‑Edit      | 7.09 | 6.76 | 6.70 |\n| **Unified Models** ||||\n| OmniGen          | 5.96 | 5.89 | 5.06 |\n| OmniGen2         | 7.16 | 6.77 | 6.41 |\n| BAGEL            | 7.36 | 6.83 | 6.52 |\n| UniWorld‑V1      | 4.93 | 7.43 | 4.85 |\n| Ovis‑U1          |   -  |   -  | 6.42 |\n| **Skywork UniPic** | **6.72** | **6.18** | **5.83** |\n\n\u003c/div\u003e\n\u003c/div\u003e\n\n## Usage\n\n### 📦 Required Packages\nCreate virtual environment and install dependencies with pip:\n```shell\nconda create -n unipic python==3.10.14\nconda activate unipic\npip install -r requirements.txt\n```\n\n\n### 📥 Checkpoints\n\nDownload the model checkpoints from [[🤗 SkyworkUniPic](https://huggingface.co/Skywork/Skywork-UniPic-1.5B)],\nIt is recommended to use the following command to download the checkpoints\n```bash\n# pip install -U \"huggingface_hub[cli]\"\nhuggingface-cli download Skywork/Skywork-UniPic-1.5B  --local-dir checkpoint --repo-type model\n```\n\n\n### 🖼️ Text-to-image Generation\n\nYou can generate images from text prompts using the following command:\n\n```shell\n# only support 1024 image_size\nexport PYTHONPATH=./:$PYTHONPATH\npython scripts/text2image.py configs/models/qwen2_5_1_5b_kl16_mar_h.py \\\n         --checkpoint checkpoint/pytorch_model.bin  --image_size 1024 \\\n         --prompt 'A glossy-coated golden retriever stands on the park lawn beside a life-sized penguin statue.'  --output output.jpg\n```\n\nTo generate a list of images based on prompts in a json file.\n```shell\nexport PYTHONPATH=./:$PYTHONPATH\naccelerate launch scripts/batch_text2image.py configs/models/qwen2_5_1_5b_kl16_mar_h.py \\\n       --checkpoint checkpoint/pytorch_model.bin  --image_size 1024 \\\n       --data data/batch_t2i.json --output output --batch_size 2 --grid_size 2\n```\nThe json file should look like:\n\n```json\n[\n  {\n   \"prompt\": \"A glossy-coated golden retriever stands on the park lawn beside a life-sized penguin statue.\"\n  },\n  {\n   \"prompt\": \"Digital portrait of a girl with rainbow hair.\"\n  }\n]\n```\n\n### ✏️ Image Editing\nThe image editing feature within this unified model is an exploratory module at the forefront of research. It currently faces challenges in edit precision, control granularity, and output consistency, and is not yet production-ready.\n\n```shell\n# only support 1024 image_size\nexport PYTHONPATH=./:$PYTHONPATH\npython scripts/image_edit.py configs/models/qwen2_5_1_5b_kl16_mar_h.py \\\n         --checkpoint checkpoint/pytorch_model.bin  --image_size 1024 \\\n         --image data/sample.png --prompt \"Replace the stars with the candle.\" \\\n         --output output.jpg \n```\n\n### 📖 Image-to-text Generation\n\n```shell\nexport PYTHONPATH=./:$PYTHONPATH\npython scripts/image2text.py configs/models/qwen2_5_1_5b_kl16_mar_h.py \\\n         --checkpoint checkpoint/pytorch_model.bin  --image_size 1024 \\\n         --image data/sample.png --prompt \"Describe the image in detail.\"\n```\n\n\n## 📜 License\nThis project is licensed under [MIT License](LICENSE).\n\n## Citation\nIf you use Skywork-UniPic in your research, please cite:\n```\n@misc{wang2025skyworkunipicunifiedautoregressive,\n      title={Skywork UniPic: Unified Autoregressive Modeling for Visual Understanding and Generation}, \n      author={Peiyu Wang and Yi Peng and Yimeng Gan and Liang Hu and Tianyidan Xie and Xiaokun Wang and Yichen Wei and Chuanxin Tang and Bo Zhu and Changshi Li and Hongyang Wei and Eric Li and Xuchen Song and Yang Liu and Yahui Zhou},\n      year={2025},\n      eprint={2508.03320},\n      archivePrefix={arXiv},\n      primaryClass={cs.CV},\n      url={https://arxiv.org/abs/2508.03320}, \n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fskyworkai%2Funipic","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fskyworkai%2Funipic","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fskyworkai%2Funipic/lists"}