Update deprecated type hinting in models (#18132)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-05-14 22:06:50 -07:00
committed by GitHub
parent 83f74c698f
commit 26d0419309
130 changed files with 971 additions and 901 deletions
+8 -7
View File
@@ -22,7 +22,8 @@
# See the License for the specific language governing permissions and
# limitations under the License.
"""Inference-only Deepseek 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
@@ -184,7 +185,7 @@ class DeepseekAttention(nn.Module):
num_heads: int,
num_kv_heads: 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,
@@ -385,8 +386,8 @@ class DeepseekModel(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"),
@@ -397,7 +398,7 @@ class DeepseekModel(nn.Module):
]
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
@@ -478,7 +479,7 @@ class DeepseekForCausalLM(nn.Module, 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]:
loader = AutoWeightsLoader(self)
return loader.load_weights(weights)