dstack icon indicating copy to clipboard operation
dstack copied to clipboard

[Roadmap] Q2 2024

Open peterschmidt85 opened this issue 1 year ago • 0 comments

This issue outlines the major items planned for Q2 2024. Note that it doesn't include bug fixes, except for major issues.

[!NOTE] Bold means priority.

Core features

  • [x] Multi-node support https://github.com/dstackai/dstack/pull/1103
  • [x] Parity YAML and CLI https://github.com/dstackai/dstack/pull/943 #1011
  • [x] More granular permissions #1101 #1138
  • [x] Support for baremetals https://github.com/dstackai/dstack/pull/1115 https://github.com/dstackai/dstack/pull/1189
  • [ ] Allow attaching volumes https://github.com/dstackai/dstack/issues/1158
  • [ ] Support private subnets #1201 #1171
  • [ ] Allow instance reuse by other users https://github.com/dstackai/dstack/issues/896
  • [ ] Scheduling tasks via CRON
  • [ ] Make sure retry policy works for all types of instances #1200

New providers

  • [ ] RunPod https://github.com/dstackai/dstack/pull/1063 https://github.com/dstackai/dstack/issues/1137
  • [ ] Oracle #1194
  • [ ] Alibaba Cloud

Provider improvements

  • [x] A10 support in Azure https://github.com/dstackai/dstack/issues/1014
  • [ ] H100 support in Azure
  • [ ] H100 support in AWS
  • [ ] H100 support in GCP
  • [ ] L4 support in AWS
  • [ ] Reserved instances in AWS https://github.com/dstackai/dstack/issues/1155
  • [ ] Reserved instances in GCP
  • [ ] Reserved instances in Azure
  • [ ] Spot instances in TensorDock
  • [ ] Spot instances in Vast.ai
  • [ ] Spot instances in RunPod

Architecture

  • [ ] TPU https://github.com/dstackai/dstack/issues/956
  • [ ] Intel Gaudi
  • [ ] Intel GPU
  • [ ] AMD GPU
  • [ ] AWS Inferentia
  • [ ] AWS Trainium

Examples

[!IMPORTANT] Community help is needed!

  • [x] H4 Alignment Handbook https://github.com/dstackai/dstack/pull/1180
  • [x] Axolotl https://github.com/dstackai/dstack/pull/1187
  • [ ] TensorRT-LLM
  • [ ] Triton
  • [ ] Function calling
  • [ ] Mixtral 8x22
  • [ ] Cog

Documentation

  • [ ] Improvements based on the feedback
  • [ ] Troubleshooting
  • [x] Best practices

[!IMPORTANT] If you notice something important missing here, please write about it in the comments.

peterschmidt85 avatar Apr 11 '24 10:04 peterschmidt85