Add block seq len and attention/activation dtype args to LLM export script #391
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
…cript Some upcomming sharded tests don't use the default values.
rsuderman
requested changes
Nov 6, 2024
| tensor_parallelism_size=tensor_parallelism_size, | ||
| use_hf=False, | ||
| static_tables=False, # Rely on the compiler for hoisting tables. | ||
| kv_cache_type="direct" if args.bs == [1] else "paged", | ||
| attention_kernel=args.attention_kernel, | ||
| attention_dtype=serialized_name_to_dtype(args.attention_dtype), |
There was a problem hiding this comment.
This depends on the other PR. It should be included there and not bleed into this PR
There was a problem hiding this comment.
I already merged the other PR.
| @@ -121,7 +121,8 @@ class LlamaModelConfig: | |||
|
|
|||
| # Block sequence stride for a paged KV cache. This must divide evenly | |||
| # into the context length. | |||
| block_seq_stride: int = 16 | |||
| default_block_seq_stride: ClassVar[int] = 16 | |||
There was a problem hiding this comment.
Why do we need ClassVar here? Seeing the default_block_seq_stride default feels bad.
There was a problem hiding this comment.
I wanted a way to reference the default from the export script, so that the CLI argument has the same default.
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Outdated suggestions cannot be applied.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.
Some upcoming sharded tests don't use the default values.