Skip to content

Conversation

xylian86
Copy link

This PR adds examples under training/DeepSpeed-SuperOffload that demonstrates full-parameter fine-tuning of popular LLMs using SuperOffload.

  • README.md — overview, quick start, NUMA/MPAM notes.
  • Training scripts for the following models:
    • GPT-OSS-20B
    • Qwen3-30B-A3B
    • Seed-OSS-36B
    • Phi-4
    • Llama-70B
    • Qwen3-14B
    • Llama-8B

Note: This PR is complimentary to PR deepspeedai/DeepSpeed#7559 on the main repo.

@xylian86 xylian86 requested a review from tjruwase as a code owner September 14, 2025 01:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant