[PP] Bypass seed checkpoint my init-ing model parts separately #516
Conversation
This file contains 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
[ghstack-poisoned]
Closed
facebook-github-bot
added
the
CLA Signed
wconstab
reviewed
Aug 12, 2024
| raise RuntimeError( | ||
| "Pipeline Parallelism requires meta-initialization and loading seed checkpoint. " | ||
| "Please run `./create_seed_checkpoint.sh` and rerun training with `--checkpoint.enable_checkpoint`" | ||
| # TODO: fix this by allowing each rank to set their own seed |
There was a problem hiding this comment.
iiuc, init_weights would be called all the time, and then when seed ckpt is specified, it would overwrite the init. that sgtm.
wconstab
approved these changes
Aug 12, 2024
…tely" Allows PP to be used without a seed checkpoint by calling `init_weight` on each model part. This is the solution in step 1 of #514 proposed by wconstab [ghstack-poisoned]
…del parts separately" Allows PP to be used without a seed checkpoint by calling `init_weight` on each model part. This is the solution in step 1 of #514 proposed by wconstab [ghstack-poisoned]
…tely" Allows PP to be used without a seed checkpoint by calling `init_weight` on each model part. This is the solution in step 1 of #514 proposed by wconstab [ghstack-poisoned]
H-Huang
added a commit
that referenced
this pull request
Aug 12, 2024
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #473 * __->__ #517 * #516 Ran `pre-commit run --all-files`
fegin
pushed a commit
that referenced
this pull request
Aug 16, 2024
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #473 * __->__ #517 * #516 Ran `pre-commit run --all-files`
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Stack from ghstack (oldest at bottom):
Allows PP to be used without a seed checkpoint by calling
init_weighton each model part. This is the solution in step 1 of #514 proposed by @wconstab