nasa-nccs-hpda/pytorch-caney

Issues

• Frontier [misc] - Add aws-olf-rccl plugin build to the training batch script

  #52 opened 4 months ago
  1

• Run plans for Frontier next steps

  #51 opened 5 months ago
  1

• Frontier [production run] - 3b - 2m - 16 patch

  #50 opened 5 months ago
  0

• Frontier [production run] - 3b - 8m - 16 patch size

  #49 opened 5 months ago
  0

• Frontier [production run] - 3b - 2m - 8 patch

  #48 opened 5 months ago
  1

• Frontier [dataset] - 8 million chip TOA dataset generation

  #47 opened 5 months ago
  0

• Frontier [planning] - Plan sequence of production runs

  #46 opened 5 months ago
  0

• Frontier [benchmarking] - 92 node scale-up benchmark

  #45 opened 5 months ago
  4

• [Frontier] Webdataset - experiencing significant IO performance on Lustre FS compared to Discover gpfs

  #44 opened 5 months ago
  1

• Frontier [benchmarking]: Benchmark with various number of nodes and batch sizes

  #43 opened 5 months ago
  1

• Frontier [testing]: swinv2-giant w/ 2 million chip dataset, 10 epochs

  #42 opened 5 months ago
  1

• Frontier [testing]: swinv2-base w/ 2 million chip dataset, 10 epochs

  #41 opened 5 months ago
  0

• Frontier [testing]: webdataset testing w/ 2 million chips

  #40 opened 5 months ago
  2

• Frontier [testing]: Webdataset small sample run

  #39 opened 5 months ago
  0

• Frontier [testing]: Run distributed pytorch benchmark

  #38 opened 5 months ago
  3

• Frontier [testing]: build conda env

  #37 opened 5 months ago
  2

• Demo MODIS-TOA for both downstream task and reconstruction loss

  #35 opened 6 months ago
  0

• Add cubo to container stack

  #34 opened 7 months ago
  0

• pt-caney architecture

  #33 opened 7 months ago
  0

• SatMAE webdataset setup

  #32 opened 7 months ago
  1

• SatMAE Integration

  #31 opened 7 months ago
  1

• Things that could need investigating (ongoing list)

  #30 opened 7 months ago
  0

• New 24.01 container version testing and including additional packages

  #29 opened 8 months ago
  0

• Class to load model weights from Hugging face similar to torchvision

  #26 opened 9 months ago
  0

• Tracking the integration of SatMAE for the 3DClouds project

  #25 opened 9 months ago
  0

• Reconsider renaming this library

  #18 opened a year ago
  0

• Tracking webdataset addition to pipeline

  #17 opened a year ago
  0

• Add Swin pre-training, model architecture into codebase

  #16 opened a year ago
  1

• Tracking horovod and torch distributed development here

  #11 opened a year ago
  0

• Refactoring to move away from model/view

  #10 opened a year ago
  3

• Implement memory-mapped dataset module

  #8 opened a year ago
  0

• Adding step by step from pytorch lightning

  #4 opened 2 years ago
  0

• Add sphinx website API

  #3 opened a year ago
  1

• Add github actions container

  #2 opened 2 years ago
  0

• Add DataLoader classes

  #1 opened a year ago
  1