[Chore] Use upstream Transformers DeepseekV4Config (#61392)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Kevin H. Luu <khluu000@gmail.com>
This commit is contained in:
Harry Mellor
2026-10-15 01:08:08 -07:00
committed by GitHub
co-authored by Claude Opus 5.5 mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Kevin H. Luu
parent 1c4c319792
commit b965ee564d
31 changed files with 155 additions and 120 deletions
@@ -31,6 +31,7 @@ from vllm.model_executor.warmup.jit_warmup_triton_helper import (
TritonWarmupTensor,
)
from vllm.platforms import current_platform
from vllm.transformers_utils.deepseek_v4_utils import get_compress_ratios
from vllm.triton_utils import tl, triton
from vllm.utils.import_utils import has_cutedsl
from vllm.utils.math_utils import cdiv, next_power_of_2
@@ -1012,7 +1013,7 @@ def _combine_topk_swa_warmup_inputs(vllm_config: Any) -> _WarmupInputRows:
"""
model_config = getattr(vllm_config, "model_config", None)
hf_config = getattr(model_config, "hf_config", None)
compress_ratios = getattr(hf_config, "compress_ratios", None) or [0]
compress_ratios = get_compress_ratios(hf_config) or [0]
topk = _hf_config_int(vllm_config, "index_topk", 0)
return zip_inputs(
*(