refactor TrainSpec to remove the name field #1850
Merged
Conversation
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
tianyu-l
requested review from
allenwang28,
ebsmothers,
fegin,
joecummings,
pbontrager,
wconstab and
wwwjn
as code owners
meta-cla
bot
added
the
CLA Signed
Contributor
|
The FLUX CI failure seems true. It seems not related to this PR, but introduced into titan long time ago. Because FLUX CI will be only triggered when you change something under flux/. I will create a PR to fix (as I'm the PoC) |
Contributor
Author
|
@wwwjn thanks! Maybe you could stamp so I can merge? |
wwwjn
approved these changes
Oct 10, 2025
ruisizhang123
pushed a commit
that referenced
this pull request
Oct 11, 2025
#1850 removed `name` field in `TrainSpec`. The experiments in simple_fsdp should also be updated. Otherwise it won't run. #1776 added `use_flex_attn` field to `apply_non_moe_tp()`, which is missing in simple_fsdp experiments ``` NGPU=8 CONFIG_FILE=./torchtitan/models/llama3/train_configs/debug_model.toml ./run_train.sh --model.name simple_fsdp.llama3 --compile.enable ``` ``` NGPU=8 CONFIG_FILE=./torchtitan/models/deepseek_v3/train_configs/debug_model.toml ./run_train.sh --model.name simple_fsdp.deepseek_v3 --compile.enable ```
githubsgi
pushed a commit
to githubsgi/torchtitan
that referenced
this pull request
Oct 13, 2025
githubsgi
pushed a commit
to githubsgi/torchtitan
that referenced
this pull request
Oct 13, 2025
pytorch#1850 removed `name` field in `TrainSpec`. The experiments in simple_fsdp should also be updated. Otherwise it won't run. pytorch#1776 added `use_flex_attn` field to `apply_non_moe_tp()`, which is missing in simple_fsdp experiments ``` NGPU=8 CONFIG_FILE=./torchtitan/models/llama3/train_configs/debug_model.toml ./run_train.sh --model.name simple_fsdp.llama3 --compile.enable ``` ``` NGPU=8 CONFIG_FILE=./torchtitan/models/deepseek_v3/train_configs/debug_model.toml ./run_train.sh --model.name simple_fsdp.deepseek_v3 --compile.enable ```
githubsgi
pushed a commit
to githubsgi/torchtitan
that referenced
this pull request
Oct 15, 2025
githubsgi
pushed a commit
to githubsgi/torchtitan
that referenced
this pull request
Oct 16, 2025
githubsgi
pushed a commit
to githubsgi/torchtitan
that referenced
this pull request
Oct 16, 2025
githubsgi
pushed a commit
to githubsgi/torchtitan
that referenced
this pull request
Oct 29, 2025
githubsgi
pushed a commit
to githubsgi/torchtitan
that referenced
this pull request
Oct 29, 2025
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.
Previously we have redundant field
job_config.model.nameandtrain_spec.nameand we used to implicitly require them to be the same.This ambiguity was exaggerated in #1740, after which we can tolerate the two fields to be different, without good reasons.
This PR removes
train_spec.name.For models inside torchtitan:
job_config.model.namewill be used to search for models in thetorchtitan/modelsfolder andtorchtitan/experimentsfolderFor models outside torchtitan:
TrainSpec, but with names explicitly specified so thatjob_config.model.namecould be used to fetch theTrainSpec