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

19
vllm/model_executor/models/grok1.py
View File

@@ -21,7 +21,8 @@
# See the License for the specific language governing permissions and
# limitations under the License.
"""Inference-only Grok1 model."""
from typing import Iterable, List, Optional, Set, Tuple, Union
from collections.abc import Iterable
from typing import Optional, Union
import torch
import torch.nn.functional as F
@@ -263,7 +264,7 @@ class Grok1DecoderLayer(nn.Module):
kv_cache: torch.Tensor,
attn_metadata: AttentionMetadata,
residual: Optional[torch.Tensor],
) -> Tuple[torch.Tensor, torch.Tensor]:
) -> tuple[torch.Tensor, torch.Tensor]:
# Self Attention
if residual is None:
residual = hidden_states
@@ -340,7 +341,7 @@ class Grok1Model(nn.Module):
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
kv_caches: List[torch.Tensor],
kv_caches: list[torch.Tensor],
attn_metadata: AttentionMetadata,
intermediate_tensors: Optional[IntermediateTensors],
inputs_embeds: Optional[torch.Tensor] = None,
@@ -371,8 +372,8 @@ class Grok1Model(nn.Module):
hidden_states, _ = self.norm(hidden_states, residual)
return hidden_states
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)
("qkv_proj", "q_proj", "q"),
@@ -390,7 +391,7 @@ class Grok1Model(nn.Module):
num_experts=num_experts)
params_dict = dict(self.named_parameters())
loaded_params: Set[str] = set()
loaded_params: set[str] = set()
for name, loaded_weight in weights:
if (self.quant_config is not None and
@@ -528,7 +529,7 @@ class Grok1ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
kv_caches: List[torch.Tensor],
kv_caches: list[torch.Tensor],
attn_metadata: AttentionMetadata,
intermediate_tensors: Optional[IntermediateTensors] = None,
inputs_embeds: Optional[torch.Tensor] = None,
@@ -547,8 +548,8 @@ class Grok1ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
sampling_metadata)
return logits
def load_weights(self, weights: Iterable[Tuple[str,
torch.Tensor]]) -> Set[str]:
def load_weights(self, weights: Iterable[tuple[str,
torch.Tensor]]) -> set[str]:
skip_prefixes = ["rotary_emb.inv_freq"]
# Skip lm_head when tie_word_embeddings is True
if self.config.tie_word_embeddings: