limoce

Posts

• FP8 is ~100 tflops faster when the kernel name has "cutlass" in it
  by limoceon 7/11/2025, 10:36 AMwith 107 comments
• Polaris: A Post-training recipe for scaling RL on Advanced Reasoning models
  by limoceon 7/9/2025, 6:58 AMwith 1 comments
• Overclocking LLM Reasoning: Monitoring and Controlling LLM Thinking Path Lengths
  by limoceon 7/6/2025, 12:53 PMwith 0 comments
• Neutrino: Probing-Based eBPF-Like GPU Kernel Profiling
  by limoceon 7/1/2025, 10:33 AMwith 0 comments
• Machine Learning Conferences Should Establish "Refutations and Critiques" Track
  by limoceon 6/26/2025, 10:27 AMwith 0 comments
• SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines
  by limoceon 3/4/2025, 7:26 AMwith 0 comments
• SepLLM: Accelerate LLMs by Compressing One Segment into One Separator
  by limoceon 3/3/2025, 1:27 PMwith 2 comments
• Step-Video-T2V: The Practice, Challenges, and Future of Video Foundation Model
  by limoceon 2/17/2025, 9:54 AMwith 5 comments
• Logic R1: Reproduce DeepSeek R1 Zero on 2K Logic Puzzle Dataset
  by limoceon 2/5/2025, 4:02 AMwith 0 comments
• Libnginx: Nginx as a Shared Library
  by limoceon 2/4/2025, 7:56 AMwith 0 comments
• DeepSeek-VL2: Moe Vision-Language Models for Advanced Multimodal Understanding [pdf]
  by limoceon 12/13/2024, 12:53 PMwith 0 comments
• Fast vectorizable algorithms of binary searching for floating point numbers
  by limoceon 11/15/2024, 12:53 AMwith 0 comments
• New OpenAI Feature: Predicted Outputs
  by limoceon 11/5/2024, 2:47 AMwith 7 comments
• Collaborative Filtering Is Wrong and Here Is Why
  by limoceon 10/24/2024, 9:03 AMwith 0 comments
• REST: A Plug-and-Play Method for Accelerating LLM Without Additional Training
  by limoceon 10/20/2024, 6:13 AMwith 0 comments
• Smoke 'em if you got 'em: Hacker gains root access using cigarette lighter
  by limoceon 10/12/2024, 1:20 PMwith 0 comments
• O1 Replication Journey: A Strategic Progress Report
  by limoceon 10/9/2024, 8:09 AMwith 0 comments
• Failures of Gradient-Based Deep Learning (2017) [pdf]
  by limoceon 8/15/2024, 10:19 AMwith 0 comments
• Qwen2-VL
  by limoceon 8/14/2024, 8:20 AMwith 0 comments
• Qwen2-Math
  by limoceon 8/8/2024, 3:00 PMwith 38 comments
• FlexAttention: The Flexibility of PyTorch with the Performance of FlashAttention
  by limoceon 8/8/2024, 7:24 AMwith 24 comments
• MiniCPM-v2.6: GPT-4V Level MLLM for Single/Multi Image and Video on Your Phone
  by limoceon 8/7/2024, 2:00 AMwith 0 comments
• MindSearch: LLM-Based Web Search Engine Similar to Perplexity.ai and SearchGPT
  by limoceon 8/1/2024, 8:53 AMwith 0 comments
• Turbo Sparse: Achieving LLM SOTA Performance with Minimal Activated Parameters
  by limoceon 6/12/2024, 11:01 AMwith 0 comments
• PowerInfer-2: Fast Large Language Model Inference on a Smartphone
  by limoceon 6/11/2024, 2:19 PMwith 0 comments
• Large-scale photonic chiplet Taichi empowers 160TOPS/W AI
  by limoceon 4/12/2024, 7:49 AMwith 0 comments
• Asterinas: OS kernel written in Rust and providing Linux-compatible ABI
  by limoceon 3/5/2024, 8:52 AMwith 0 comments
• Mq-deadline scalability improvements (with more than 100% improvement)
  by limoceon 1/20/2024, 12:04 PMwith 0 comments