Spec v2 tree drafting (topk>1) with page_size>1 (#26972)
This commit is contained in:
@@ -87,11 +87,32 @@ patches:
|
||||
"""
|
||||
|
||||
|
||||
_PR_REVERT_YAML_26972 = """
|
||||
patches:
|
||||
- target: sglang.srt.model_executor.model_runner_kv_cache_mixin.ModelRunnerKVCacheMixin._init_pools
|
||||
edits:
|
||||
- match: |
|
||||
if (
|
||||
self.server_args.speculative_algorithm is not None
|
||||
and self.server_args.page_size > 1
|
||||
and (self.server_args.speculative_eagle_topk or 1) > 1
|
||||
):
|
||||
from sglang.srt.managers.utils import get_alloc_len_per_decode
|
||||
|
||||
extra_max_context_len = max(
|
||||
extra_max_context_len,
|
||||
2 * get_alloc_len_per_decode(self.server_args),
|
||||
)
|
||||
replacement: ""
|
||||
"""
|
||||
|
||||
|
||||
_PR_FIX_REVERT_YAML: Dict[int, str] = {
|
||||
25015: _PR_REVERT_YAML_25015,
|
||||
26329: _PR_REVERT_YAML_26329,
|
||||
27338: _PR_REVERT_YAML_27338,
|
||||
27360: _PR_REVERT_YAML_27360,
|
||||
26972: _PR_REVERT_YAML_26972,
|
||||
}
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user