UltraBERT: One Encoder for Classification and Retrieval
dmis-lab
• • 59
NEW Articles from Team or Enterprise organizations will get promoted to the main section.
UltraBERT: One Encoder for Classification and Retrieval
dmis-lab
• • 59The Missing Piece in LLM Memory: Why Long Context and Retrieval Aren't Enough for Belief Revision
boltwave
• • 57Tokenmaxxing BERT: 50x faster tokenization
huggingface
• • 49Introducing Lumma-Fev-27B
FrontiersMind
• • 26One Decision Model Instead of a Ranking Team: Short-Video and News Recommendation with autotrust/JEV-27B-VL
autotrust
• • 22SlideOn & mSlideOn: Sparsified Late-Interaction and Dense Retrievers
lightonai
• • 16When agents self-improve, what should persist?
BunsDev
• • 16Impactful scheduling for GPU clusters
allenai
• • 54Backpropagation on an Inference-Only NPU: Pretraining VKORE-46M-NPU From Scratch
FINAL-Bench
• • 13Fine-Tune ASR for Your Domain with NVIDIA Speech Agent Skills
nvidia
• • 12ColGrep-Agent: Cutting Frontier Coding Agents’ Tokens by a Third
lightonai
• • 20LightOnOCR-3: High-Performance OCR and Layout Extraction in One Model
lightonai
• • 39Sort your prompts: a free speed-up for MoE batch inference
badaoui
• • 8The Most Intuitive Way to Understand Score Function Centering: Understand Trainer-Generator Mismatch in Async RL from the Score Function
NormalUhr
• • 8Estimating physics from images: what real measurements reveal
projectsim
• • 8QVAC TranscribePsy: State-of-the-Art On-Device Multilingual Speech Recognition
qvac
• • 7New in llama.cpp: Decision Models
ggml-org
• • 105Introducing Llama Windows: llama.cpp in your Windows tray
huggingface
• • 10Extending Nemotron 3.5 Super 120B VL with Agentic Web Development Capabilities
tngtech
• • 7NVIDIA Kumo Tabular Sets a New Accuracy-Efficiency Frontier for Tabular Prediction
nvidia
• • 96