Fix bad synthetic dataloader with per device batch size < 1. #1706
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
wang2yn84
requested review from
gobbleturk,
khatwanimohit,
bvandermoon,
vipannalla,
RissyRan,
richjames0,
rni418,
gagika,
shralex,
yangyuwei,
SurbhiJainUSC,
hengtaoguo,
A9isha,
wyzhang,
mitalisi,
gpolovets1,
mailvijayasingh,
jrplatin,
patemotter and
Lumosis
as code owners
wang2yn84
changed the title
bvandermoon
reviewed
May 9, 2025
| @@ -83,6 +83,7 @@ def __init__(self, config, mesh): | |||
| self.mesh = mesh | |||
| dataset = BadSyntheticDataIterator.get_bad_synthetic_data(config) | |||
| self.data_generator = multihost_dataloading.MultiHostDataLoadIterator(dataset, self.mesh) | |||
| self.local_iterator = self.data_generator.local_iterator | |||
There was a problem hiding this comment.
Does it need to be used somewhere?
There was a problem hiding this comment.
Hi Branden, yes it's used here
Line 193 in 9b5fca5
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Outdated suggestions cannot be applied.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.
Description
Fix BadSyntheticDataIterator for grain. The local iterator is missing and workload will error out on when using Grain dataset together with pdb < 1.
Tests
Manually run the following workload: python -m MaxText.train MaxText/configs/base.yml skip_jax_distributed_system=True run_name=lance_test attention=dot_product dataset_type=grain tokenizer_path=assets/tokenizer.llama2 hardware=gpu logits_dot_in_fp32=false enable_goodput_recording=false monitor_goodput=false remat_policy=full weight_dtype=bfloat16 save_config_to_gcs=false scan_layers=false per_device_batch_size=0.25 dcn_fsdp_parallelism=-1 dcn_data_parallelism=1 ici_fsdp_parallelism=1 ici_tensor_parallelism=8 packing=false enable_checkpoint_cloud_logger=true dataset_path=/scratch/lancewang/dataset_pvc/ grain_train_files=/scratch/lancewang/dataset_pvc/array-record/c4/en/3.0.1/c4-train.array_record* grain_worker_count=1 enable_checkpointing=false async_checkpointing=true checkpoint_period=10 save_config_to_gcs=false base_output_directory=/scratch/lancewang/outputs
Checklist
Before submitting this PR, please make sure (put X in square brackets):