Skip to content

Add support for vLLM KV-cache quantization #773

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 2 commits into
base: main
Choose a base branch
from
Open

Add support for vLLM KV-cache quantization #773

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d00a7c4
Add support for vLLM KV-cache quantization
eldarkurtic May 22, 2025
e97c5b9
Merge branch 'main' into patch-2
NathanHB May 22, 2025
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: 4 additions & 1 deletion src/lighteval/models/vllm/vllm_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,9 +104,10 @@ class VLLMModelConfig(ModelConfig):
max_num_seqs: PositiveInt = 128 # maximum number of sequences per iteration; This variable and `max_num_batched_tokens` effectively control the batch size at prefill stage. See https://github.com/vllm-project/vllm/issues/2492 for detailed explaination.
max_num_batched_tokens: PositiveInt = 2048 # maximum number of tokens per batch
subfolder: str | None = None
kv_cache_dtype: str = "auto"
calculate_kv_scales: bool = False
is_async: bool = False # Whether to use the async version or sync version of the model


class VLLMModel(LightevalModel):
def __init__(
self,
Expand Down Expand Up @@ -187,6 +188,8 @@ def _create_auto_model(self, config: VLLMModelConfig) -> Optional[LLM]:
"seed": int(config.seed),
"max_num_seqs": int(config.max_num_seqs),
"max_num_batched_tokens": int(config.max_num_batched_tokens),
"kv_cache_dtype": config.kv_cache_dtype,
"calculate_kv_scales": config.calculate_kv_scales,
}

if config.quantization is not None:
Expand Down
Loading