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

Actual fix (#9787)

Browse Source
This commit is contained in:
Lysandre Debut
2021-01-25 17:12:07 +01:00
committed by GitHub
parent fac7cfb16a
commit 0f443436fb
2 changed files with 15 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/transformers/models/gpt2/modeling_gpt2.py
View File

@@ -541,6 +541,7 @@ class GPT2Model(GPT2PreTrainedModel):
self.ln_f = nn.LayerNorm(config.n_embd, eps=config.layer_norm_epsilon) self.ln_f = nn.LayerNorm(config.n_embd, eps=config.layer_norm_epsilon)
self.init_weights() self.init_weights()
# Model parallel # Model parallel
self.model_parallel = False self.model_parallel = False
self.device_map = None self.device_map = None
@@ -805,7 +806,9 @@ class GPT2LMHeadModel(GPT2PreTrainedModel):
self.init_weights() self.init_weights()
# Model parallel
self.model_parallel = False self.model_parallel = False
self.device_map = None
@add_start_docstrings(PARALLELIZE_DOCSTRING) @add_start_docstrings(PARALLELIZE_DOCSTRING)
def parallelize(self, device_map=None): def parallelize(self, device_map=None):
@@ -971,6 +974,10 @@ class GPT2DoubleHeadsModel(GPT2PreTrainedModel):
self.init_weights() self.init_weights()
# Model parallel
self.model_parallel = False
self.device_map = None
def get_output_embeddings(self): def get_output_embeddings(self):
return self.lm_head return self.lm_head
@@ -1153,6 +1160,10 @@ class GPT2ForSequenceClassification(GPT2PreTrainedModel):
self.init_weights() self.init_weights()
# Model parallel
self.model_parallel = False
self.device_map = None
@add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING) @add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
@add_code_sample_docstrings( @add_code_sample_docstrings(
tokenizer_class=_TOKENIZER_FOR_DOC, tokenizer_class=_TOKENIZER_FOR_DOC,

4
src/transformers/models/t5/modeling_t5.py
View File

@@ -1651,6 +1651,10 @@ class T5EncoderModel(T5PreTrainedModel):
self.init_weights() self.init_weights()
# Model parallel
self.model_parallel = False
self.device_map = None
@add_start_docstrings(PARALLELIZE_DOCSTRING) @add_start_docstrings(PARALLELIZE_DOCSTRING)
def parallelize(self, device_map=None): def parallelize(self, device_map=None):
self.device_map = ( self.device_map = (