SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Use Python 3.9 syntax in examples (#37279)

Browse Source 
Signed-off-by: cyy <cyyever@outlook.com>
This commit is contained in:
cyyever
2025-04-07 19:52:21 +08:00
committed by GitHub
parent 08f36771b3
commit 0fb8d49e88
123 changed files with 358 additions and 451 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
examples/modular-transformers/modeling_multimodal2.py
View File

@@ -5,7 +5,7 @@
# modular_multimodal2.py file directly. One of our CI enforces this.
# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
from typing import Optional, Tuple, Union
from typing import Optional, Union
import torch
from torch import nn
@@ -65,7 +65,7 @@ class Multimodal2VisionAttention(nn.Module):
attention_mask: Optional[torch.Tensor] = None,
causal_attention_mask: Optional[torch.Tensor] = None,
output_attentions: Optional[bool] = False,
) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
) -> tuple[torch.Tensor, Optional[torch.Tensor]]:
"""Input shape: Batch x Time x Channel"""
bsz, tgt_len, embed_dim = hidden_states.size()
@@ -152,7 +152,7 @@ class Multimodal2VisionSdpaAttention(Multimodal2VisionAttention):
attention_mask: Optional[torch.Tensor] = None,
causal_attention_mask: Optional[torch.Tensor] = None,
output_attentions: Optional[bool] = False,
) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
) -> tuple[torch.Tensor, Optional[torch.Tensor]]:
if output_attentions:
# TODO: Improve this warning with e.g. `model.config.attn_implementation = "manual"` once this is implemented.
logger.warning_once(
@@ -233,7 +233,7 @@ class Multimodal2VisionFlashAttention2(Multimodal2VisionAttention):
attention_mask: Optional[torch.Tensor] = None,
causal_attention_mask: Optional[torch.Tensor] = None,
output_attentions: Optional[bool] = False,
) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
) -> tuple[torch.Tensor, Optional[torch.Tensor]]:
output_attentions = False
batch_size, q_len, _ = hidden_states.size()
@@ -334,7 +334,7 @@ class Multimodal2VisionEncoderLayer(nn.Module):
attention_mask: torch.Tensor,
causal_attention_mask: torch.Tensor,
output_attentions: Optional[bool] = False,
) -> Tuple[torch.FloatTensor]:
) -> tuple[torch.FloatTensor]:
"""
Args:
hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
@@ -392,7 +392,7 @@ class Multimodal2VisionEncoder(nn.Module):
output_attentions: Optional[bool] = None,
output_hidden_states: Optional[bool] = None,
return_dict: Optional[bool] = None,
) -> Union[Tuple, BaseModelOutput]:
) -> Union[tuple, BaseModelOutput]:
r"""
Args:
inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
@@ -587,7 +587,7 @@ class Multimodal2VisionTransformer(nn.Module):
output_hidden_states: Optional[bool] = None,
return_dict: Optional[bool] = None,
interpolate_pos_encoding: Optional[bool] = False,
) -> Union[Tuple, BaseModelOutputWithPooling]:
) -> Union[tuple, BaseModelOutputWithPooling]:
r"""
Returns:
@@ -671,7 +671,7 @@ class Multimodal2VisionModel(Multimodal2VisionPreTrainedModel):
output_hidden_states: Optional[bool] = None,
interpolate_pos_encoding: bool = False,
return_dict: Optional[bool] = None,
) -> Union[Tuple, BaseModelOutputWithPooling]:
) -> Union[tuple, BaseModelOutputWithPooling]:
r"""
Returns: