Reading list for research in generation models.
We list the most popular methods for generation models, if we missed something, please submit a request. (Note: We show the date the first edition of the paper was submitted to arxiv, but the link to the paper may be up to date.)
Backbone:
| Date | Method | Conference | Title | Code/Project Page | abstract |
|---|---|---|---|---|---|
| 2024 | VisionLLaMA | ECCV 2024 | VisionLLaMA: A Unified LLaMA Backbone for Vision Tasks | code | 对视觉语言大模型的大一统backbone进行了探讨,主要提出了一个2d位置编码 |
Autoregressive generation model:
Diffusion model:
Segmentation:
| Date | Method | Conference | Title | Code |
|---|---|---|---|---|
| 2021-12-06 | ddpm-segmentation | ICLR 2022 | Label-Efficient Semantic Segmentation with Diffusion Models | ddpm-segmentation |
Other Discriminative Tasks:
| Date | Method | Conference | Title | Code |
|---|---|---|---|---|
| 2023-05-18 | ~ | Arxiv 2023 | Discriminative Diffusion Models as Few-shot Vision and Language Learners | None |
Survey:
| Date | Conference | Title |
|---|---|---|
| 2022-09-02 | Arxiv 2022 | Diffusion Models: A Comprehensive Survey of Methods and Applications |
| 2022-09-10 | Arxiv 2022 | Diffusion Models in Vision: A Survey |
| 2022-09-12 | Arxiv 2022 | A Survey on Generative Diffusion Model |
| 2023-04-02 | Arxiv 2023 | Text-to-image Diffusion Models in Generative AI:A Survey |
| Sep 2024 | AI 2024 | Multi-Modal Generative AI: Multi-modal LLM, Diffusion and Beyond |
Large Language Models:
| Date | Method | Conference | Title | Code |
|---|---|---|---|---|
| Feb 2023 | LLAMA | Arxiv Feb 2023 | Llama: Open and efficient foundation language models | code |
Model Merging:
HFRL:
| Date | Method | Conference | Title | Code/Project Page | abstract |
|---|---|---|---|---|---|
| 6 Oct 2024 | TIS-DPO | arxiv 6 Oct 2024 | TIS-DPO: Token-level Importance Sampling for Direct Preference Optimization With Estimated Weights | [None] | 正在看,别催啦 |
| 17 Oct 2024 | TIS-DPO | arxiv 17 Oct 2024 | Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein Design | code | 提出了一种RL算法,结合蛋白质领域的离散扩散方法,直接通过奖励对模型进行优化解决不可导问题 |
Tutorial:
CVPR 2022 Tutorial:Denoising Diffusion-based Generative Modeling: Foundations and Applications