Skip to content

Clean up unused variables in get_kvcache_metadata_buffer #371

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
gameofdimension wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Clean up unused variables in get_kvcache_metadata_buffer #371

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
5 changes: 0 additions & 5 deletions examples/hstu/configs/inference_config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -262,16 +262,11 @@ def get_kvcache_metadata_buffer(
hstu_config: InferenceHSTUConfig, kvcache_config: KVCacheConfig
):
device = torch.cuda.current_device()
torch.bfloat16 if hstu_config.bf16 else torch.float16 if hstu_config.fp16 else torch.float32

max_new_history_seqlen = hstu_config.max_batch_size * hstu_config.max_seq_len
max_num_pages_per_seq = (
hstu_config.max_seq_len + hstu_config.max_seq_len + kvcache_config.page_size - 1
) // kvcache_config.page_size
max_host_kv_buffer_size = (
hstu_config.max_batch_size * hstu_config.max_seq_len,
hstu_config.num_heads * hstu_config.head_dim,
)

default_num_pages_per_seq = 4
paged_indices_buffer = torch.randint(
Expand Down