Egor Dmitriev

Recent Posts

  • CoreML, Partitions, and the 13x Mac Speedup

    May 18, 2026

  • CUDA, TensorRT, and the FP16 Softmax Overflow

    May 17, 2026

  • From Research Checkpoint to ONNX Runtime

    May 16, 2026

  • Building Boosters: A Gradient Boosting Library from Scratch

    Feb 03, 2026

See 15 more →

Tag: cuda

1 item with this tag.

  • May 17, 2026

    CUDA, TensorRT, and the FP16 Softmax Overflow

    • deep-learning
    • computer-vision
    • tensorrt
    • optimization
    • onnx
    • cuda

  • GitHub
  • LinkedIn
  • RSS Feed