论文阅读
论文阅读
扩散模型
- 扩散模型 DDPM — Denoising Diffusion Probabilistic Models
- 注意力机制 Transformer — Attention Is All You Need
- 隐式扩散模型 DDIM — Denoising Diffusion Implicit Models
- 无分类器引导 CFG — Classifier-Free Diffusion Guidance
- 潜空间扩散模型 LDM — High-Resolution Image Synthesis with Latent Diffusion Models
- 条件控制 ControlNet — Adding Conditional Control to Text-to-Image Diffusion Models
- 扩散 Transformer DiT — Scalable Diffusion Models with Transformers
- 流匹配 Flow Matching — Flow Matching for Generative Modeling
- 整流流 Rectified Flow — Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
- 多模态扩散 Transformer SD3 — Scaling Rectified Flow Transformers for High-Resolution Image Synthesis
大语言模型
- 注意力机制 Transformer — Attention Is All You Need
- 生成式预训练 GPT — Improving Language Understanding by Generative Pre-Training
- 上下文学习 GPT-3 — Language Models are Few-Shot Learners
- 训练规模与数据 Chinchilla — Training Compute-Optimal Large Language Models
- 基础语言模型 LLaMA — LLaMA: Open and Efficient Foundation Language Models
- 低秩微调 LoRA — LoRA: Low-Rank Adaptation of Large Language Models
- 指令微调与对齐 InstructGPT — Training language models to follow instructions with human feedback
- 直接偏好优化 DPO — Direct Preference Optimization: Your Language Model is Secretly a Reward Model
- 检索增强生成 RAG — Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- 思维链 CoT — Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
- 推理强化学习 DeepSeek-R1 — DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning