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 10, 00:15 UTC
[Misc] Remove unused ModelKeys (#32608)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Jee Jee Li committed
Jan 20, 2026 at 01:34 UTC
2636d76257af5f1bbdb42bcf7d05680be2a1f625
1 file changed
+1
-38
vllm/model_executor/models/module_mapping.py
+1
-38
@@ -8,44 +8,7 @@ from dataclasses import dataclass, field
8
9
10
@dataclass
11
-class ModelKeys:
12
- model_type: str = None
13
-
14
- module_list: str = None
15
-
16
- embedding: str = None
17
-
18
- mlp: str = None
19
-
20
- down_proj: str = None
21
-
22
- attention: str = None
23
-
24
- o_proj: str = None
25
-
26
- q_proj: str = None
27
-
28
- k_proj: str = None
29
-
30
- v_proj: str = None
31
-
32
- qkv_proj: str = None
33
-
34
- qk_proj: str = None
35
-
36
- qa_proj: str = None
37
-
38
- qb_proj: str = None
39
-
40
- kva_proj: str = None
41
-
42
- kvb_proj: str = None
43
-
44
- output: str = None
45
-
46
-
47
-@dataclass
48
-class MultiModelKeys(ModelKeys):
11
+class MultiModelKeys:
12
language_model: list[str] = field(default_factory=list)
13
connector: list[str] = field(default_factory=list)
14
# vision tower and audio tower