biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Support LoRA and GPTQModel for PLaMo 2/3 (#31322)

Browse Source 
Signed-off-by: Shinichi Hemmi <50256998+Alnusjaponica@users.noreply.github.com>
This commit is contained in:
Shinichi Hemmi
2025-12-26 12:41:33 +09:00
committed by GitHub
parent 3b8f31b362
commit 9ee05cbe7f
3 changed files with 31 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
vllm/model_executor/models/plamo3.py
View File

@@ -35,7 +35,7 @@ from vllm.model_executor.model_loader.weight_utils import (
composed_weight_loader,
default_weight_loader,
)
from vllm.model_executor.models.interfaces import SupportsPP
from vllm.model_executor.models.interfaces import SupportsLoRA, SupportsPP
from vllm.model_executor.models.utils import (
AutoWeightsLoader,
extract_layer_index,
@@ -369,13 +369,10 @@ class Plamo3Model(nn.Module):
return hidden_states
class Plamo3ForCausalLM(nn.Module, SupportsPP):
class Plamo3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
packed_modules_mapping = {
"qkv_proj": [
"q_proj",
"k_proj",
"v_proj",
],
"qkv_proj": ["qkv_proj"],
"gate_up_proj": ["gate_up_proj"],
}
def __init__(self, *, vllm_config: VllmConfig, prefix: str = "") -> None: