Daily Papers of 2024-04-25

1. CatLIP: CLIP-level Visual Recognition Accuracy with 2.7x Faster Pre-training on Web-scale Image-Text Data 22 upvotes, #1 of 2024-04-25
2. PuLID: Pure and Lightning ID Customization via Contrastive Alignment 15 upvotes, #2 of 2024-04-25
3. ID-Aligner: Enhancing Identity-Preserving Text-to-Image Generation with Reward Feedback Learning 11 upvotes, #3 of 2024-04-25
4. MoDE: CLIP Data Experts via Clustering 11 upvotes, #3 of 2024-04-25
5. MotionMaster: Training-free Camera Motion Transfer For Video Generation 10 upvotes, #5 of 2024-04-25
6. Editable Image Elements for Controllable Synthesis 9 upvotes, #6 of 2024-04-25
7. MaGGIe: Masked Guided Gradual Human Instance Matting 8 upvotes, #7 of 2024-04-25
8. XC-Cache: Cross-Attending to Cached Context for Efficient LLM Inference 6 upvotes, #8 of 2024-04-25
9. BASS: Batched Attention-optimized Speculative Sampling 5 upvotes, #9 of 2024-04-25

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.