A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 12:15 UTC
Remove duplicate code from dbrx.py (#17550)
sstamenk committed
May 1, 2025 at 20:51 UTC
04f2cfc89455a05bbeca18c9d5e6c55cd146c560
1 file changed
+1
-5
vllm/model_executor/models/dbrx.py
+1
-5
@@ -451,13 +451,9 @@ class DbrxForCausalLM(nn.Module, SupportsPP):
451
break
452
453
else:
454
- # Remapping the name of FP8 kv-scale.
455
- name = maybe_remap_kv_scale_name(name, params_dict)
456
- if name is None:
457
- continue
458
-
454
if is_pp_missing_parameter(name, self):
455
continue
456
+ # Remapping the name of FP8 kv-scale.
457
name = maybe_remap_kv_scale_name(name, params_dict)
458
if name is None:
459
continue