HN Explorer

antinucleon

↗ HN profile [ 60.6 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. The Great Wave Has Arrived (Memo from GLM CEO Jie Tang) (twitter.com)
    1 points 0
  2. What Muse Spark 1.1 Taught Us About Enterprise Agent Architecture (int21.ai)
    2 points 0
  3. Stop Waiting for a Bigger Context Window (int21.ai)
    3 points 1
  4. Show HN: INT21 – Self-Improving PTX Kernel Factory (int21.ai)
    3 points 0
  5. 7 days of autonomous agent search outperformed FlashAttention-4 and CUDNN (twitter.com)
    3 points 0
  6. From Node.js/Python to PTX: The first AI framework generated by AI agents (github.com)
    3 points 0
  7. PetaFLOPS Inference Era: 1 Pflops Attention, and Preliminary End-to-End Results (blog.hippoml.com)
    2 points 0
  8. Unified DataCenter and Local Foundation Model Serving: Beyond Docker Way (blog.hippoml.com)
    4 points 0
  9. Super AI Creativity App Run with Local GPU on Windows/Linux/MacOS (blog.hippoml.com)
    19 points 10
  10. Learning to Optimize Tensor Programs (arxiv.org)
    106 points 13
  11. Build Your Own TensorFlow with NNVM and Torch (dmlc.ml)
    4 points 0
  12. Deep Learning in Scala (dmlc.ml)
    2 points 0
  13. MXNetJS: JavaScript Package for Deep Learning in Browser (without Server) (github.com)
    1 points 0
  14. Deep Learning in a Single Source File for Smart Devices (mxnet.readthedocs.org)
    2 points 0
  15. Training Deep Net on 14M Images by Using a Single Machine (mxnet.readthedocs.org)
    6 points 0
  16. Programming Models for Deep Learning (mxnet.readthedocs.org)
    4 points 0
  17. Remove and Ban Cheaters Who Sent Multi-Petition Before H1B Lottery Selection (petitions.whitehouse.gov)
    4 points 0
  18. Why we still need expression template in C++ (github.com)

    In wikipedia, it is said by using rvalue reference in C++11, we don't need expression template. But in machine learning software especially with CUDA, expression template is still a powerful tool.

    2 points 0
  19. MSHADOW: Lightweight Matrix and Tensor Template Library in C++/CUDA (github.com)

    MSHADOW is a lightweight C++ tensor expression system support frequently used functions in machine learning. By using MSHADOW, all code can be written once and run smoothly on both GPU and CPU.

    1 points 0
  20. CXXNET: A Deep Conv NN with 1000 lines core code (github.com)

    CXXNET is yet another implementation of (convolutional) neural network. It is in C++, with about 1000 lines of network layer implementations, it is able to train ImageNet in fast speed.

    2 points 0