feat: track checkpoint progress in micro-batches - #195
Open
JYMiracle305 wants to merge 2 commits into
Open
Conversation
JYMiracle305
force-pushed
the
feat/checkpoint-consumed-micro-batches
branch
from
August 3, 2026 08:27
49d8d4b to
57ab888
Compare
JYMiracle305
force-pushed
the
feat/checkpoint-consumed-micro-batches
branch
from
August 4, 2026 07:43
57ab888 to
7689011
Compare
Chamberlain0w0
requested changes
Aug 5, 2026
| struct TrainerState { | ||
| int64_t global_step = 0; | ||
| int64_t consumed_batches = 0; | ||
| int64_t consumed_micro_batches = 0; |
Contributor
There was a problem hiding this comment.
目前 ckpt 保存的是 DP × 每 rank 已处理 micro-batch 数，没有包含 batch_size。当恢复时改变 --batch_size，非 PP 路径会跳过相同数量的 micro-batch，却对应不同数量的样本；PP 路径还可能在整除检查处退出，或者跳到错误位置。建议参考 Megatron 风格的 consumed_train_samples 记法，或者加一个 FIXME，暂时提醒一下禁止 resume 时修改 bs
Contributor
Author
There was a problem hiding this comment.
将 ckpt 内记录的值改为 consumed_train_samples，现在支持 resume 时修改 bs 了
Contributor
Author
精度对比
性能对比
|
JYMiracle305
force-pushed
the
feat/checkpoint-consumed-micro-batches
branch
2 times, most recently
from
August 6, 2026 08:44
59a9674 to
9d99035
Compare
JYMiracle305
force-pushed
the
feat/checkpoint-consumed-micro-batches
branch
from
August 6, 2026 09:47
9d99035 to
862a378
Compare
kilinchange
force-pushed
the
feat/checkpoint-consumed-micro-batches
branch
from
August 7, 2026 02:18
862a378 to
f6f4715
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Depends on #194
主要修改
该 PR 统一了 checkpoint 中训练进度的计量语义，将原先含义模糊的 consumed_batches 重命名为 consumed_micro_batches。
在不同流水线并行配置下，DataLoader 单次迭代包含的 micro-batch 数量不同。该 PR 将 checkpoint 中保存的进度统一为 micro-batch 数量，并在恢复训练时根据当前 PP 配置换算为对应的 DataLoader 位置，同时正确处理 DDP rank 和步长。