Skip to content

fea(readme): VLLM_PROMPT_SEQ_BUCKET_MAX value update #2121

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: v1.23.0_next
Choose a base branch
from
Open

fea(readme): VLLM_PROMPT_SEQ_BUCKET_MAX value update #2121

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions README_GAUDI.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -405,7 +405,7 @@ batch size is often at its maximum, making large-batch HPU graphs critical to ca
- batch size max (`VLLM_PROMPT_BS_BUCKET_MAX`): `min(max_num_seqs, 64)`
- sequence length min (`VLLM_PROMPT_SEQ_BUCKET_MIN`): `block_size`
- sequence length step (`VLLM_PROMPT_SEQ_BUCKET_STEP`): `block_size`
- sequence length max (`VLLM_PROMPT_SEQ_BUCKET_MAX`): `max_model_len`
- sequence length max (`VLLM_PROMPT_SEQ_BUCKET_MAX`): `input_tokens` (or `max_model_len` if maximum input token size is not known)

- Decode:

Expand All @@ -418,7 +418,7 @@ batch size is often at its maximum, making large-batch HPU graphs critical to ca
- Recommended Values:
- Prompt:

- sequence length max (`VLLM_PROMPT_SEQ_BUCKET_MAX`): `input_tokens + output_tokens` rounded up to a multiple of `block_size` (especially recommended for models with high max_model_len)
- sequence length max (`VLLM_PROMPT_SEQ_BUCKET_MAX`): `input_tokens + output_tokens` rounded up to a multiple of `block_size` (especially recommended for models with high `max_model_len`)
- Decode:

- block size max (`VLLM_DECODE_BLOCK_BUCKET_MAX`): `max(128, (max_num_seqs*max_model_len/block_size)`
Expand Down