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

Add LoRA support for DeepSeek models (V2, V3, R1-0528) (#23971)

Browse Source 
Signed-off-by: sadeghja1070 <sadegh.ja1070@gmail.com>
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
This commit is contained in:
sadegh.shokatian
2025-08-30 06:40:02 -07:00
committed by GitHub
parent 3a6acad431
commit 379ea2823a
3 changed files with 12 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
vllm/model_executor/models/deepseek.py
View File

@@ -52,7 +52,7 @@ from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.sequence import IntermediateTensors
from .interfaces import SupportsPP
from .interfaces import SupportsLoRA, SupportsPP
from .utils import (AutoWeightsLoader, extract_layer_index,
is_pp_missing_parameter,
make_empty_intermediate_tensors_factory, make_layers,
@@ -439,7 +439,11 @@ class DeepseekModel(nn.Module):
return loaded_params
class DeepseekForCausalLM(nn.Module, SupportsPP):
class DeepseekForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
packed_modules_mapping = {
"qkv_proj": ["q_proj", "k_proj", "v_proj"],
"gate_up_proj": ["gate_proj", "up_proj"],
}
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
super().__init__()

5
vllm/model_executor/models/deepseek_v2.py
View File

@@ -56,7 +56,7 @@ from vllm.model_executor.model_loader.weight_utils import (
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.sequence import IntermediateTensors
from .interfaces import MixtureOfExperts, SupportsPP
from .interfaces import MixtureOfExperts, SupportsLoRA, SupportsPP
from .utils import (PPMissingLayer, is_pp_missing_parameter,
make_empty_intermediate_tensors_factory, make_layers,
maybe_prefix)
@@ -727,7 +727,8 @@ class DeepseekV2Model(nn.Module):
return hidden_states
class DeepseekV2ForCausalLM(nn.Module, SupportsPP, MixtureOfExperts):
class DeepseekV2ForCausalLM(nn.Module, SupportsPP, MixtureOfExperts,
SupportsLoRA):
packed_modules_mapping = {
"gate_up_proj": ["gate_proj", "up_proj"],
}