[Core] Support dynamically loading Lora adapter from HuggingFace (#6234)

Co-authored-by: Antoni Baum <antoni.baum@protonmail.com>
This commit is contained in:
Jiaxin Shan
2024-07-22 15:42:40 -07:00
committed by GitHub
co-authored by Antoni Baum
parent 69d5ae38dc
commit 42c7f66a38
11 changed files with 201 additions and 18 deletions
+1 -1
View File
@@ -691,7 +691,7 @@ class GPUModelRunnerBase(ModelRunnerBase[TModelInputForGPU]):
dummy_lora_request = LoRARequest(
lora_name=f"warmup_{lora_id}",
lora_int_id=lora_id,
lora_local_path="/not/a/real/path",
lora_path="/not/a/real/path",
)
self.lora_manager.add_dummy_lora(dummy_lora_request,
rank=LORA_WARMUP_RANK)