论文阅读 6
- SeqDiffuSeq, Text Diffusion with Encoder-Decoder Transformers
- DiffuSeq, Sequence to Sequence Text Generation with Diffusion Models
- Diffusion-LM Improves Controllable Text Generation
- Disentangling Reasoning Capabilities from Language Models with Compositional Reasoning Transformers
- The Power of Scale for Parameter-Efficient Prompt Tuning
- Prefix-Tuning, Optimizing Continuous Prompts for Generation