[misc] fix: gradient accumulation in seq balance and modify default vllm log level #141
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
| @@ -125,6 +125,7 @@ def __init__(self, config: DictConfig, role: str): | |||
| self.config.actor.ppo_micro_batch_size //= (self.device_mesh.shape[0] // | |||
| self.ulysses_sequence_parallel_size) | |||
| self.config.actor.ppo_micro_batch_size_per_gpu = self.config.actor.ppo_micro_batch_size | |||
| assert self.config.actor.ppo_mini_batch_size % self.config.actor.ppo_micro_batch_size_per_gpu == 0 | |||
There was a problem hiding this comment.
We also have to encore self.config.actor.ppo_mini_batch_size >= n_gpus * self.config.actor.ppo_micro_batch_size_per_gpu
There was a problem hiding this comment.
Is it necessary? The mini_batch_size here is already normalized, then if self.config.actor.ppo_mini_batch_size < n_gpus * self.config.actor.ppo_micro_batch_size_per_gpu, the above line will not get 0
vermouth1992
approved these changes
Jan 27, 2025
Chendong98
pushed a commit
to Chendong98/verl
that referenced
this pull request
Feb 4, 2025
…llm log level (volcengine#141) - Previous gradient accumulation value is computed by micro_batch_size, which is wrong when using dynamic_bsz - Fix ci script to avoid overlooking this issue - Change vLLM state log default value to True to disable log. - We will check the `self.config.actor.ppo_mini_batch_size % self.config.actor.ppo_micro_batch_size_per_gpu == 0` after normalization in fsdp_workers instead of in dp_actor and dp_critic.
as12138
pushed a commit
to as12138/verl
that referenced
this pull request
Feb 20, 2025
…llm log level (volcengine#141) - Previous gradient accumulation value is computed by micro_batch_size, which is wrong when using dynamic_bsz - Fix ci script to avoid overlooking this issue - Change vLLM state log default value to True to disable log. - We will check the `self.config.actor.ppo_mini_batch_size % self.config.actor.ppo_micro_batch_size_per_gpu == 0` after normalization in fsdp_workers instead of in dp_actor and dp_critic.
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.
self.config.actor.ppo_mini_batch_size % self.config.actor.ppo_micro_batch_size_per_gpu == 0after normalization in fsdp_workers instead of in dp_actor and dp_critic.