Mirrors/modelscope
1
0
Fork 0
mirror of https://github.com/modelscope/modelscope.git synced 2026-09-01 19:49:03 +02:00
Code Issues Packages Projects Releases Wiki Activity

[to #43115513] bug fix for nlp backbone-head trainers

Browse Source 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/9662281
This commit is contained in:
zhangzhicheng.zzc
2022-08-06 13:55:46 +08:00
committed by wenmeng.zwm
parent aef3fa34bd
commit ceee95f763
2 changed files with 31 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

33
modelscope/models/nlp/backbones/structbert.py
View File

@@ -31,24 +31,23 @@ class SbertModel(TorchModel, SbertModelTransform):
def extract_pooled_outputs(self, outputs):
return outputs['pooler_output']
def forward(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_values=None,
use_cache=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
):
def forward(self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_values=None,
use_cache=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
**kwargs):
return SbertModelTransform.forward(
self, input_ids, attention_mask, token_type_ids, position_ids,
head_mask, inputs_embeds, encoder_hidden_states,
encoder_attention_mask, past_key_values, use_cache,
output_attentions, output_hidden_states, return_dict)
output_attentions, output_hidden_states, return_dict, **kwargs)

31
modelscope/models/nlp/structbert/modeling_sbert.py
View File

@@ -870,22 +870,21 @@ class SbertModel(SbertPreTrainedModel):
output_type=BaseModelOutputWithPoolingAndCrossAttentions,
config_class=_CONFIG_FOR_DOC,
)
def forward(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_values=None,
use_cache=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
):
def forward(self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_values=None,
use_cache=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
**kwargs):
r"""
encoder_hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length, hidden_size)`,
`optional`):