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

Update deprecated type hinting in models (#18132)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-05-15 06:06:50 +01:00
committed by GitHub
parent 83f74c698f
commit 26d0419309
130 changed files with 971 additions and 901 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
vllm/model_executor/models/deepseek_v2.py
View File

@@ -22,7 +22,8 @@
# See the License for the specific language governing permissions and
# limitations under the License.
"""Inference-only DeepseekV2/DeepseekV3 model."""
from typing import Any, Dict, Iterable, Optional, Set, Tuple, Union
from collections.abc import Iterable
from typing import Any, Optional, Union
import torch
from torch import nn
@@ -200,7 +201,7 @@ class DeepseekV2Attention(nn.Module):
q_lora_rank: int,
kv_lora_rank: int,
rope_theta: float = 10000,
rope_scaling: Optional[Dict[str, Any]] = None,
rope_scaling: Optional[dict[str, Any]] = None,
max_position_embeddings: int = 8192,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
@@ -352,7 +353,7 @@ class DeepseekV2MLAAttention(nn.Module):
q_lora_rank: Optional[int],
kv_lora_rank: int,
rope_theta: float = 10000,
rope_scaling: Optional[Dict[str, Any]] = None,
rope_scaling: Optional[dict[str, Any]] = None,
max_position_embeddings: int = 8192,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
@@ -736,8 +737,8 @@ class DeepseekV2ForCausalLM(nn.Module, SupportsPP):
device=device),
})
def load_weights(self, weights: Iterable[Tuple[str,
torch.Tensor]]) -> Set[str]:
def load_weights(self, weights: Iterable[tuple[str,
torch.Tensor]]) -> set[str]:
stacked_params_mapping = [
# (param_name, shard_name, shard_id)
("gate_up_proj", "gate_proj", 0),
@@ -753,7 +754,7 @@ class DeepseekV2ForCausalLM(nn.Module, SupportsPP):
num_experts=self.config.n_routed_experts)
params_dict = dict(self.named_parameters())
loaded_params: Set[str] = set()
loaded_params: set[str] = set()
for name, loaded_weight in weights:
if "rotary_emb.inv_freq" in name:
continue