Skip to content

[1/15] Add true-on-policy contract policy surface - #34

Open
maocheng23 wants to merge 19 commits into
miles-mainfrom
split/pr30-01-contracts
Open

[1/15] Add true-on-policy contract policy surface#34
maocheng23 wants to merge 19 commits into
miles-mainfrom
split/pr30-01-contracts

Conversation

@maocheng23

Copy link
Copy Markdown

Stacked split of #30.

Base branch: miles-main
Head branch: split/pr30-01-contracts

This is PR 1 of 15. The stack is cumulative; the final branch matches the original PR head.

yueming-yuan and others added 19 commits February 25, 2026 19:22
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

Co-authored-by: Yueming Yuan <yym022502@gmail.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
- Detach output layer params to prevent MTP gradient flowing to output layer
- Add mtp_kwargs interface for flexible MTP label/loss_mask passing
- Roll mtp_labels and loss_mask for RL training compatibility

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Yueming Yuan <yym022502@gmail.com>
…yers (#10)

- Add is_mtp flag to MoE layers and multi_token_prediction module
- Bypass routing replay for MTP layers (MTP uses fresh routing)
- Replace rdxa/dev's built-in RouterReplay with miles.utils.routing_replay:
  - moe_utils.py: use get_routing_replay_compute_topk() wrapper
  - router.py: use register_routing_replay() for initialization

Co-authored-by: Yueming Yuan <yym022502@gmail.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Yueming Yuan <yym022502@gmail.com>
After bumping Megatron (rdxa/dev), colocated IPC weight update fails with
torch.AcceleratorError: CUDA error: invalid argument during
torch.multiprocessing serialization of CUDA tensors.

Root cause: Megatron's new TMS hook (PR NVIDIA#3048) alters allocator behavior
in training flow, causing allocations via cuMemCreate/cuMemMap which are
incompatible with CUDA IPC (_share_cuda_() fails).

Fix: resolve mapping.py and dynamic_context.py conflicts to isolate
hook side effects so TMS/allocator state remains IPC-compatible during
the weight update phase.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- merge(): truncate dp_reshardable padding on optimizer/param_state path
- load_parameter_state_from_dp_reshardable: tolerate missing 'padding' key
- ShardedTensor: relax flattened_range to deprecation warning

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This PR rebases from radixark Megatron fork [miles-20260218](https://github.com/radixark/Megatron-LM/tree/miles-20260218) and resolve conflicts.

Upgrade Megatron from Dec 17 (3714d81) to Feb 13 (1dcf0da)

PR link: #13

Co-authored-by: Yueming Yuan <yym022502@gmail.com>
Made-with: Cursor
…se `--disable-weight-backuper` in miles (#18)

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: fzyzcjy <5236035+fzyzcjy@users.noreply.github.com>
Co-authored-by: fzyzcjy <5236035+fzyzcjy@users.noreply.github.com>
Squash merge of the dense true-on-policy Megatron branch.

Co-authored-by: zju-stu-lizheng <lizheng.cs@zju.edu.cn>
Co-authored-by: zyxiyy02 <282300612+zyxiyy02@users.noreply.github.com>
Co-authored-by: Yi Zhang <1109276519@qq.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@maocheng23
maocheng23 marked this pull request as ready for review May 26, 2026 22:45
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.

4 participants