What's Papers with Code?
Papers with Code is a platform that highlights trending research in the field of machine learning along with the code to implement it. It serves as a comprehensive resource for researchers and developers looking to stay updated with the latest advancements and practical implementations in machine learning.
Benefits of Papers with Code
- Access to Latest Research: Stay updated with the most recent papers in machine learning.
- Implementation Ready: Direct access to code implementations, facilitating experimentation and learning.
- Benchmarking: Compare the performance of different models on a variety of tasks.
- Community Engagement: Engage with a community of researchers and developers sharing insights and improvements.
How to Use Papers with Code
- Explore Papers: Browse through a curated list of machine learning papers sorted by relevance and impact.
- Access Code: Directly access repositories containing the code implementations of the research.
- Benchmark Models: View and compare model performances across standard datasets and tasks.
- Contribute: Share improvements or new implementations with the community.
Featured Papers
MossFormer: Pushing the Performance Limit of Monaural Speech Separation
- Description: Utilizes a gated single-head transformer with convolution-augmented joint self-attentions to improve speech separation.
- Code Repository: Available on GitHub under
modelscope/ClearerVoice-Studio.
Gaze-LLE: Gaze Target Estimation via Large-Scale Learned Encoders
- Description: Focuses on predicting gaze targets in a scene using learned encoders.
- Code Repository: Available on GitHub under
fkryan/gazelle.
Segment Any Text: A Universal Approach for Sentence Segmentation
- Description: Introduces the SaT model for robust and efficient sentence segmentation.
- Code Repository: Available on GitHub under
facebookresearch/large_concept_model.
StableAnimator: High-Quality Identity-Preserving Human Image Animation
- Description: Enhances face quality during animation using Hamilton-Jacobi-Bellman equation-based optimization.
- Code Repository: Available on GitHub under
Francis-Rings/StableAnimator.
SynCamMaster: Synchronizing Multi-Camera Video Generation
- Description: Advances in video diffusion models for consistent multi-camera video generation.
- Code Repository: Available on GitHub under
kwaivgi/syncammaster.
Learning Flow Fields in Attention for Controllable Person Image Generation
- Description: Improves performance of diffusion models using a model-agnostic loss.
- Code Repository: Available on GitHub under
franciszzj/leffa.
Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation
- Description: Transitions single-image 3D human modeling to a data-centric paradigm.
- Code Repository: Available on GitHub under
isarandi/nlf.
Video Seal: Open and Efficient Video Watermarking
- Description: Introduces a framework for neural video watermarking.
- Code Repository: Available on GitHub under
facebookresearch/videoseal.
HunyuanVideo: A Systematic Framework For Large Video Generative Models
- Description: An open-source model demonstrating high performance in video generation.
- Code Repository: Available on GitHub under
tencent/hunyuanvideo.
LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods
- Description: Analyzes limitations of LLM judges and future directions.
- Code Repository: Available on GitHub under
cshaitao/awesome-llms-as-judges.