Model accuracy alignment#2
Open
zhengshengning wants to merge 5 commits into
Open
Conversation
This comment was marked as outdated.
This comment was marked as outdated.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
问题2: 禁用 TE 配置 (ms-swift前置修改)
问题描述: 为了使拆分后的独立 LayerNorm 模块正常工作，需要禁用 Transformer Engine 融合算子。
修复文件: ms-swift-test/swift/megatron/model/register.py
修改PR：modelscope@175dd87
问题3: 不添加任何后缀token (ms-swift前置修改)
修复文件: swift/template/templates/utils.py
修改PR：modelscope@bd1726c
问题10: MG 梯度走 fp32 累加
查配置确认：configs/minimax_v25_single_torch.yaml 和 run_torch_minimax.sh 都没设 accumulate_allreduce_grads_in_fp32，所以 MG 的 main_grad 是 bf16，喂给 AdamW 的是 bf16 梯度；而 PF amp_master_grad: true 喂的是 fp32。step1 前反向全对齐（bf16 计算一致），但 optimizer 拿到的梯度精度不同 → 权重更新不同 → step2 前向从 embedding 起分叉