Milestones

List view

• support apple silicon gpu

  No due date
  0% complete0 open0 closed

• implement scalable and efficient tensor code structure

  - separate the runtime implementation from cranberry/tensor.py - manage Tensor.grad as another Tensor, not a NumPy ndarray - etc

  Overdue by 2 year(s)

  •

  Due by August 31, 2024
  0% complete0 open0 closed

• run representative models using Cranberry

  - mnist: https://towardsdatascience.com/going-beyond-99-mnist-handwritten-digits-recognition-cfff96337392 - resnet-50: this is important because we will use it for mlperf - gpt-2: https://youtu.be/l8pRSuU81PU?si=p_x5-DNH5qp7ZjHT - efficientnet - yolov8 - mnist gan - llama - stable diffusion (sorted by highest priority)

  No due date

  •1/1 issues closed
  100% complete0 open1 closed

• support nvidia gpu

  No due date
  0% complete0 open0 closed