32B工业代码模型首发，战争验证推理真伪

重点关注

• InCoder-32B: Code Foundation Model for Industrial Scenarios score 10
  入选 HF Daily Papers；HF 热度: 286 upvotes (+4)；有代码实现；关键词(4): post-training, pre-training, reasoning, open-source
• MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild score 10
  入选 HF Daily Papers；HF 热度: 97 upvotes (+4)；有代码实现；关键词(2): production, fine-tuning
• MosaicMem: Hybrid Spatial Memory for Controllable Video World Models score 10
  机构: Amazon；入选 HF Daily Papers；HF 热度: 72 upvotes (+4)
• Kinema4D: Kinematic 4D World Modeling for Spatiotemporal Embodied Simulation score 10
  入选 HF Daily Papers；HF 热度: 66 upvotes (+4)；有代码实现；关键词(1): embodied
• Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training score 10
  入选 HF Daily Papers；HF 热度: 28 upvotes (+4)；有代码实现；关键词(1): pre-training
• When AI Navigates the Fog of War score 10
  入选 HF Daily Papers；HF 热度: 22 upvotes (+4)；有代码实现；关键词(1): reasoning
• Omnilingual MT: Machine Translation for 1,600 Languages score 10
  机构: Meta；入选 HF Daily Papers；HF 热度: 12 upvotes (+3)；关键词(1): leaderboard
• SocialOmni: Benchmarking Audio-Visual Social Interactivity in Omni Models score 9
  入选 HF Daily Papers；HF 热度: 102 upvotes (+4)；有代码实现
• WorldCam: Interactive Autoregressive 3D Gaming Worlds with Camera Pose as a Unifying Geometric Representation score 9
  入选 HF Daily Papers；HF 热度: 53 upvotes (+4)；有代码实现
• SegviGen: Repurposing 3D Generative Model for Part Segmentation score 9
  入选 HF Daily Papers；HF 热度: 16 upvotes (+3)；有代码实现；关键词(1): distillation

也值得关注

• Anticipatory Planning for Multimodal AI Agents score 4
  入选 HF Daily Papers；关键词(2): fine-tuning, reasoning
• AI Scientist via Synthetic Task Scaling score 4
  入选 HF Daily Papers；关键词(3): scaling, agentic, code generation
• Parallel In-context Learning for Large Vision Language Models score 4
  关键词(2): latency, vision-language；顶会接收: CVPR
• Pre-training LLM without Learning Rate Decay Enhances Supervised Fine-Tuning score 4
  关键词(2): fine-tuning, pre-training；顶会接收: ICLR
• Boosting Quantitive and Spatial Awareness for Zero-Shot Object Counting score 4
  关键词(1): reasoning；顶会接收: CVPR
• Adaptive Theory of Mind for LLM-based Multi-Agent Coordination score 4
  关键词(1): reasoning；顶会接收: AAAI
• Locate-then-Sparsify: Attribution Guided Sparse Strategy for Visual Hallucination Mitigation score 4
  关键词(2): deployment, vision-language；顶会接收: CVPR
• ProgressiveAvatars: Progressive Animatable 3D Gaussian Avatars score 4
  关键词(1): real-time；顶会接收: CVPR
• Face2Scene: Using Facial Degradation as an Oracle for Diffusion-Based Scene Restoration score 4
  关键词(1): compression；顶会接收: CVPR
• Omnilingual SONAR: Cross-Lingual and Cross-Modal Sentence Embeddings Bridging Massively Multilingual Text and Speech score 4
  机构: Meta；关键词(1): distillation
• ACPV-Net: All-Class Polygonal Vectorization for Seamless Vector Map Generation from Aerial Imagery score 4
  关键词(1): edge；顶会接收: CVPR
• Kestrel: Grounding Self-Refinement for LVLM Hallucination Mitigation score 4
  机构: Princeton；关键词(3): deployment, tool use, vision-language
• $x^2$-Fusion: Cross-Modality and Cross-Dimension Flow Estimation in Event Edge Space score 4
  关键词(1): edge；顶会接收: CVPR
• When Should a Robot Think? Resource-Aware Reasoning via Reinforcement Learning for Embodied Robotic Decision-Making score 4
  机构: Harvard；关键词(3): latency, reasoning, embodied
• Probing Cultural Signals in Large Language Models through Author Profiling score 4
  机构: INRIA；关键词(2): fine-tuning, open-source