The official GitHub page for the survey paper "CLIP-Powered Domain Generalization and Domain Adaptation: A Comprehensive Survey". This paper has been accepted for publication in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI).

1. 2021_ICML_CLIP_Learning Transferable Visual Models From Natural Language Supervision.

   [ICML] [arXiv] [GitHub]

1. Introduction

2. Preliminaries

3. Domain Generalization

3.1 Prompt Optimization Techniques

1. 2022_CVPR_CoCoOp_Conditional Prompt Learning for Vision-Language Models.

   [CVPR] [arXiv] [GitHub]

2. 2022_IJCV_CoOp_Learning to Prompt for Vision-Language Models.

   [ACM] [Springer] [arXiv] [GitHub]

3. 2023_CVPR_KgCoOp_Visual-Language Prompt Tuning with Knowledge-Guided Context Optimization.

   [CVPR] [arXiv] [GitHub]

4. 2023_CVPR_MaPLe_MaPLe: Multi-Modal Prompt Learning.

   [CVPR] [IEEE] [arXiv] [GitHub]

5. 2023_ICCV_ProGrad_Prompt-aligned Gradient for Prompt Tuning.

   [ICCV] [IEEE] [arXiv] [GitHub]

6. 2024_AAAI_LAMM_LAMM: Label Alignment for Multi-Modal Prompt Learning.

   [AAAI] [arXiv] [GitHub]

3.2 CLIP is Adopted as Backbone or Encoder

3.2.1 Source-Available (SA)

3.2.1.1 Single-Source Closed-Set Domain Generalization (SS-CSDG)

(i) prompt-driven optimization methods.

1. 2023_ICCV_DAPT_Distribution-Aware Prompt Tuning for Vision-Language Models.

   [ICCV] [arXiv] [GitHub] [GitHub 2]

2. 2023_ICCV_PromptSRC_Self-Regulating Prompts: Foundational Model Adaptation without Forgetting.

   [ICCV] [IEEE] [GitHub]

3. 2024_ECCV_GalLoP_GalLoP: Learning Global and Local Prompts for Vision-Language Models.

   [ECCV] [arXiv] [GitHub]

4. 2024_arXiv_LDFS_Enhancing Vision-Language Models Generalization via Diversity-Driven Novel Feature Synthesis.

   [arXiv]

5. 2024_ECCV_SPG_Soft Prompt Generation for Domain Generalization.

   [ECCV] [arXiv] [GitHub]

6. 2025_CVPR Workshop_FrogDogNet_FrogDogNet: Fourier frequency Retained visual prompt Output Guidance for Domain Generalization of CLIP in Remote Sensing.

   [CVPR] [arXiv] [GitHub]

(ii) architecture-enhanced alignment methods.

1. 2023_ICCV_BorLan_Borrowing Knowledge From Pre-trained Language Model: A New Data-efficient Visual Learning Paradigm.

   [ICCV] [IEEE] [GitHub]

2. 2024_CVPR_MMA_MMA: Multi-Modal Adapter for Vision-Language Models.

   [CVPR] [IEEE] [GitHub]

3. 2024_WACV_StyLIP_StyLIP: Multi-Scale Style-Conditioned Prompt Learning for CLIP-based Domain Generalization.

   [WACV] [arXiv]

3.2.1.2 Multi-Source Closed-Set Domain Generalization (MS-CSDG)

(i) Distillation-based methods

1. 2023_ICCV_RISE_A Sentence Speaks a Thousand Images: Domain Generalization through Distilling CLIP with Language Guidance.

   [ICCV] [IEEE] [arXiv]

2. 2024_CVPR_VL2V-ADiP_Leveraging Vision-Language Models for Improving Domain Generalization in Image Classification.

   [CVPR] [IEEE] [arXiv] [Homepage] [GitHub]

3. 2024_Access_CAL_Consistent Augmentation Learning for Generalizing CLIP to Unseen Domains.

   [IEEE]

(ii) Prompt-driven methods

1. 2023_TJSAI_DPL_Domain Prompt Learning for Efficiently Adapting CLIP to Unseen Domains.

   [Jstage] [GitHub]

2. 2024_arXiv_SPG_Soft Prompt Generation for Domain Generalization.

   [ECCV] [arXiv] [GitHub]

3. 2024_CVPR_Any-Shift Prompting_Any-Shift Prompting for Generalization over Distributions.

   [CVPR] [IEEE] [arXiv] [GitHub]

4. 2024_CVPR_DPR_Disentangled Prompt Representation for Domain Generalization.

   [CVPR] [IEEE]

5. 2025_arXiv_PADG_Prompt Disentanglement via Language Guidance and Representation Alignment for Domain Generalization.

   [arXiv]

6. 2025_CVPR_Diverse-Text-Prompts_Domain Generalization in CLIP via Learing with Diverse Text Prompts.

   [CVPR] [IEEE]

(iii) Architecture-level Enhancement methods

1. 2024_AIEA_Mixup CLIPood_Robust Domain Generalization for Multi-modal Object Recognition.

   [IEEE] [arXiv]

2. 2024_CVPR_ODG-CLIP_Unknown Prompt, the only Lacuna: Unveiling CLIP’s Potential for Open Domain Generalization.

   [CVPR] [IEEE] [arXiv] [GitHub]

3. 2024_WACV_StyLIP_StyLIP: Multi-Scale Style-Conditioned Prompt Learning for CLIP-based Domain Generalization.

   [WACV] [IEEE] [arXiv]

4. 2025_arXiv_HAM_Harmonizing and Merging Source Models for CLIP-based Domain Generalization.

   [arXiv]

5. 2025_TCSVT_ClipMix_ClipMix for Domain Generalization.

   [IEEE]

6. 2025_WACV_MoA_Domain Generalization using Large Pretrained Models with Mixture-of-Adapters.

   [IEEE] [arXiv] [HomePage] [GitHub]

3.2.1.3 Multi-Source Open-Set Domain Generalization (MS-OSDG)

1. 2023_ICML_CLIPood_CLIPood: Generalizing CLIP to Out-of-Distributions.

   [ICML] [GitHub]

2. 2024_CVPR_ODG-CLIP_Unknown Prompt, the only Lacuna: Unveiling CLIP’s Potential for Open Domain Generalization.

   [CVPR] [GitHub]

3. 2024_CVPR_SCI-PD_PracticalDG: Perturbation Distillation on Vision-Language Models for Hybrid Domain Generalization.

   [CVPR] [GitHub]

4. 2025_CVPR_OSLoPrompt_OSLoPrompt: Bridging Low Supervision Challenges and Open-Set Domain Generalization in CLIP.

   [CVPR] [arXiv] [GitHub]

5. 2025_TMLR_MetaPrompt_Meta-Learning to Teach Semantic Prompts for Open Domain Generalization in Vision-Language Models.

   [OpenReview]

3.2.2 Source-Free (SF)

Source(-Fully)-Free Domain Generalization (S(F)F-DG)

1. 2022_arXiv_DUPRG_Domain-Unified Prompt Representations for Source-Free Domain Generalization.

   [arXiv] [GitHub]

2. 2023_ICCV_PromptStyler_PromptStyler: Prompt-driven Style Generation for Source-free Domain Generalization.

   [ICCV] [IEEE] [arXiv] [GitHub]

3. 2024_arXiv_2025_ToM_DPStyler_DPStyler: Dynamic PromptStyler for Source-Free Domain Generalization.

   [arXiv] [ToM] [ACM] [GitHub]

4. 2024_arXiv_2025_ICASSP_PromptTA_PromptTA: Prompt-driven Text Adapter for Source-free Domain Generalization.

   [arXiv] [ICASSP] [GitHub]

5. 2025_TCSVT_BatStyler_BatStyler= Advancing Multi-category Style Generation for Source-free Domain Generalization.

   [IEEE] [arXiv] [GitHub]

4. Domain Adaptation

4.1 Source-Available (SA)

4.1.1 Single-Source (SS)

4.1.1.1 Single-Source Closed-Set Unsupervised Domain Adaptation (SS-CSUDA)

(i) Domain-Aware and Dual-Branch Prompt Tuning.

1. 2023_ICCV_PADCLIP_PADCLIP: Pseudo-labeling with Adaptive Debiasing in CLIP for Unsupervised Domain Adaptation.

   [ICCV] [IEEE]

2. 2023_ICCVW_AD-CLIP_AD-CLIP: Adapting Domains in Prompt Space Using CLIP.

   [ICCV] [IEEE] [arXiv] [GitHub]

3. 2023_TNNLS_DAPrompt_Domain Adaptation via Prompt Learning.

   [IEEE] [arXiv] [GitHub]

4. 2024_AAAI_PDA_Prompt-based Distribution Alignment for Unsupervised Domain Adaptation.

   [AAAI] [arXiv] [GitHub]

5. 2024_WACV_PTT-VFR_Empowering Unsupervised Domain Adaptation with Large-scale Pre-trained Vision-Language Models.

   [WACV] [IEEE]

6. 2025_IJCV_MAwLLM_Multi-modal Prompt Alignment with Fine-grained LLM Knowledge for Unsupervised Domain Adaptation.

   [Springer]

7. 2025_KBS_PDbDa_PDbDa: Prompt-Tuned Dual-Branch Framework for Unsupervised Domain Adaptation.

   [Elsevier]

8. 2025_TIP_ADAPT_When Adversarial Training Meets Prompt Tuning: Adversarial Dual Prompt Tuning for Unsupervised Domain Adaptation.

   [ACM] [IEEE] [GitHub]

(ii) Cross-Modal Alignment and Bridging Mechanisms.

1. 2024_CVPR_DAMP_Domain-Agnostic Mutual Prompting for Unsupervised Domain Adaptation.

   [CVPR] [IEEE] [arXiv] [GitHub]

2. 2024_CVPR_UniMoS_Split to Merge: Unifying Separated Modalities for Unsupervised Domain Adaptation.

   [CVPR] [IEEE] [arXiv] [GitHub]

3. 2025_TGRS_PADA-Net_Prompt-Integrated Adversarial Unsupervised Domain Adaptation for Scene Recognition.

   [IEEE]

4. 2025_TOMM_PIMA_Prompt-Based Invertible Mapping Alignment for Unsupervised Domain Adaptation.

   [ACM]

5. 2025_TOMM_UTISA_Unified Text-Image Space Alignment with Cross-Modal Prompting in CLIP for UDA.

   [ACM]

(iii) Distribution Alignment and Regularization Strategies.

1. 2024_arXiv_CLIP-Div_CLIP the Divergence: Language-guided Unsupervised Domain Adaptation.

   [arXiv]

2. 2024_IJCNN_DACR_CLIP-Enhanced Unsupervised Domain Adaptation with Consistency Regularization.

   [IEEE]

3. 2024_TCSVT_CMKD_Unsupervised Domain Adaption Harnessing Vision-Language Pre-training.

   [IEEE] [ACM] [arXiv] [GitHub]

4. 2025_WACV_SWG_Combining Inherent Knowledge of Vision-Language Models with Unsupervised Domain Adaptation through Strong-weak Guidance.

   [IEEE] [arXiv]

(iv) Fuzzy System–Enhanced Adaptation.

1. 2024_FUZZ-IEEE_VLM-TSK-DA_Enhancing Vision-Language Models Incorporating TSK Fuzzy System for Domain Adaptation.

   [IEEE]

2. 2024_TOFS_FUZZLE_Unsupervised Domain Adaptation Enhanced by Fuzzy Prompt Learning.

   [IEEE] [ACM]

4.1.1.2 Single-Source Open-Set Unsupervised Domain Adaptation (SS-OSUDA)

1. 2023_arXiv_2025_CVIU_ODA with CLIP_Open-Set Domain Adaptation with Visual-Language Foundation Models.

   [ScienceDirect CVIU] [ACM] [arXiv]

2. 2024_arXiv_COSMo_COSMo: CLIP Talks on Open-Set Multi-Target Domain Adaptation.

   [arXiv] [GitHub]

3. 2024_ICIP_PromptIDIV_Decoupling Domain Invariance and Variance With Tailored Prompts for Open-Set Domain Adaptation.

   [IEEE]

4. 2025_CVIU_entropy optimization_Open-set Domain Adaptation with Visual-Language Foundation Models.

   [ScienceDirect] [arXiv]

4.1.2 Multi-Source (MS)

4.1.2.1 Multi-Source Closed-Set Unsupervised Domain Adaptation (MS-CSUDA)

1. 2023_NeurIPS_MPA_Multi-Prompt Alignment for Multi-Source Unsupervised Domain Adaptation.

   [NeurIPS] [ACM] [arXiv] [GitHub]

2. 2024_arXiv_LanDA_LanDA: Language-Guided Multi-Source Domain Adaptation.

   [arXiv] [GitHub]

3. 2024_OpenReview_MSDPL_Domain Prompt Matters a Lot in Multi-Source Few-Shot Domain Adaptation.

   [OpenReview]

4. 2025_AAAI_VAMP_Vision-aware Multimodal Prompt Tuning for Uploadable Multi-source Few-Shot Domain Adaptation.

   [AAAI] [arXiv] [GitHub]

5. 2025_CVPR_CRPL_Preserving Clusters in Prompt Learning for Unsupervised Domain Adaptation.

   [CVPR] [arXiv]

4.1.2.2 Multi-Source Open-Partial-Set Unsupervised Domain Adaptation (MS-OPSUDA) a.k.a. Universal Multi-Source Domain Adaptation (UniMDA)

1. 2024_Signal Processing Letters_SAP-CLIP_Semantic-Aware Adaptive Prompt Learning for Universal Multi-Source Domain Adaptation.

   [IEEE]

4.2 Source-Free (SF)

4.2.1 Source-Fully-Free (SFF)

4.2.1.1 Source-Fully-Free Closed-Set Unsupervised Domain Adaptation (SFF-CSUDA) a.k.a. Unsupervised Fine-Tuning (CS-UFT)

1. 2022_arXiv_UPL_Unsupervised Prompt Learning for Vision-Language Models.

   [arXiv] [GitHub]

2. 2023_ICML_POUF_POUF: Prompt-oriented unsupervised fine-tuning for large pre-trained models.

   [PMLR] [ACM] [arXiv] [GitHub]

3. 2024_arXiv_2025_WACV_DPA_DPA: Dual Prototypes Alignment for Unsupervised Adaptation of Vision-Language Models.

   [WACV] [IEEE] [arXiv] [GitHub]

4. 2024_arXiv_TFUP_Training-Free Unsupervised Prompt for Vision-Language Models.

   [arXiv] [GitHub]

5. 2024_ICIP_Rethinking Domain Adaptation and Generalization in the Era Of CLIP.

   [IEEE] [arXiv]

6. 2024_ICML_CPL_Candidate Pseudolabel Learning: Enhancing Vision-Language Models by Prompt Tuning with Unlabeled Data.

   [PMLR] [ACM] [arXiv] [GitHub] [Slides]

7. 2024_ICML_UEO_Realistic Unsupervised CLIP Fine-tuning with Universal Entropy Optimization.

   [PMLR] [ACM] [arXiv] [GitHub]

8. 2024_NeurIPS_LaFTer_LaFTer: Label-Free Tuning of Zero-shot Classifier using Language and Unlabeled Image Collections.

   [NeurIPS] [ACM] [arXiv] [Homepage] [GitHub]

9. 2024_WACV_ReCLIP_ReCLIP: Refine Contrastive Language Image Pre-Training with Source Free Domain Adaptation.

   [WACV] [IEEE] [arXiv] [GitHub]

4.2.1.2 Source-Fully-Free Partial-Set Unsupervised Domain Adaptation (SFF-PSUDA) a.k.a. Partial-Set Unsupervised Fine-Tuning (PS-UFT)

1. 2024_ICML_UEO_Realistic Unsupervised CLIP Fine-tuning with Universal Entropy Optimization.

   [PMLR] [ACM] [arXiv] [GitHub]

4.2.1.3 Source-Fully-Free Open-Set Unsupervised Domain Adaptation (SFF-OSUDA) a.k.a. Open-Set Unsupervised Fine-Tuning (OS-UFT)

1. 2023_ICML Workshop_UOTA_UOTA: Unsupervised Open-Set Task Adaptation Using a Vision-Language Foundation Model.

   [ICML]

2. 2024_ICML_UEO_Realistic Unsupervised CLIP Fine-tuning with Universal Entropy Optimization.

   [PMLR] [ACM] [arXiv] [GitHub]

3. 2025_arXiv_CLIPXpert_Revisiting CLIP for SF-OSDA= Unleashing Zero-Shot Potential with Adaptive Threshold and Training-Free Feature Filtering.

   [arXiv]

4.2.1.4 Source-Fully-Free Open-Partial-Set Unsupervised Domain Adaptation (SFF-OPSUDA) a.k.a Open-Partial-Set Unsupervised Fine-Tuning (OPS-UFT)

1. 2024_ICML_UEO_Realistic Unsupervised CLIP Fine-tuning with Universal Entropy Optimization.

   [PMLR] [ACM] [arXiv] [GitHub]

4.2.2 Source-Data-Free (SDF)

4.2.2.1 Source-Data-Free Closed-Set Unsupervised Domain Adaptation (SDF-CSUDA)

1. 2024_arXiv_CDBN_Data-Efficient CLIP-Powered Dual-Branch Networks for Source-Free Unsupervised Domain Adaptation.

   [arXiv] [GitHub]

2. 2024_CVPR_DIFO_Source-Free Domain Adaptation with Frozen Multimodal Foundation Model.

   [CVPR] [IEEE] [arXiv] [GitHub]

3. 2024_IJCV_Co-learn++_Source-Free Domain Adaptation Guided by Vision and Vision-Language Pre-Training.

   [Springer] [ACM] [arXiv] [GitHub]

4. 2024_Signal-Image-and-Video-Processing_BBC_CLIP-guided Black-Box Domain Adaptation of Image Classification.

   [Springer] [OpenReview]

5. 2025_ICLR_Prode_Proxy Denoising for Source-Free Domain Adaptation.

   [OpenReview] [arXiv] [Slides] [GitHub]

6. 2025_NeurIPS_DUET_DUET: Dual-Perspective Pseudo Labeling and Uncertainty-aware Exploration & Exploitation Training for Source-Free Domain Adaptation.

   [OpenReview] [GitHub]

4.2.2.2 Source-Data-Free Partial-Set Unsupervised Domain Adaptation (SDF-PSUDA)

1. 2024_CVPR_DIFO_Source-Free Domain Adaptation with Frozen Multimodal Foundation Model.

   [CVPR] [IEEE] [arXiv] [GitHub]

2. 2024_IJCV_Co-learn++_Source-Free Domain Adaptation Guided by Vision and Vision-Language Pre-Training.

   [Springer] [ACM] [arXiv] [GitHub]

4.2.2.3 Source-Data-Free Open-Set Unsupervised Domain Adaptation (SDF-OSUDA)

1. 2023_arXiv_2025_CVIU_ODA with CLIP_Open-Set Domain Adaptation with Visual-Language Foundation Models.

   [ScienceDirect CVIU] [ACM] [arXiv]

2. 2024_CVPR_DIFO_Source-Free Domain Adaptation with Frozen Multimodal Foundation Model.

   [CVPR] [IEEE] [arXiv] [GitHub]

3. 2024_IJCV_Co-learn++_Source-Free Domain Adaptation Guided by Vision and Vision-Language Pre-Training.

   [Springer] [ACM] [arXiv] [GitHub]

Datasets and Metrics

Common Datasets

Related Survey

1. 2024_AIReview_Survey_Domain Generalization through Meta-Learning: A Survey.

   [Springer] [arXiv] [d-bn.info]

2. 2022_TKDE_Survey_Generalizing to Unseen Domains: A Survey on Domain Generalization. [IEEE] [IJCAI] [arXiv] [Slides]

3. 2024_TPAMI_Survey_A Comprehensive Survey on Source-Free Domain Adaptation.

   [TPAMI]

4. 2024_NN_Survey_Source-free unsupervised domain adaptation: A survey.

   [ScienceDirect] [ACM] [arXiv]

5. 2024_IJCV_Survey_A Comprehensive Survey on Test-Time Adaptation Under Distribution Shifts.

   [IJCV] [arXiv] [GitHub]

Related Reposority

1. DomainBed.

   [GitHub]

2. tim-learn/awesome-test-time-adaptation

   [GitHub]

📖 Citation

If you find this work helpful, please consider citing our paper:

@article{li2025clip,
  title={CLIP-Powered Domain Generalization and Domain Adaptation: A Comprehensive Survey},
  author={Li, Jindong and Li, Yongguang and Fu, Yali and Liu, Jiahong and Liu, Yixin and Yang, Menglin and King, Irwin},
  journal={arXiv preprint arXiv:2504.14280},
  year={2025}
}