Skip to content

Generate - update cookie cutters to not initialize cache with training and gradient checkpointing #21759

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
gante merged 1 commit into from
Feb 24, 2023
Merged

Generate - update cookie cutters to not initialize cache with training and gradient checkpointing #21759

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 13 additions & 15 deletions ...tter-template-{{cookiecutter.modelname}}/modeling_{{cookiecutter.lowercase_modelname}}.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -184,7 +184,7 @@ def forward(
token_type_ids = buffered_token_type_ids_expanded
else:
token_type_ids = torch.zeros(input_shape, dtype=torch.long, device=self.position_ids.device)

if inputs_embeds is None:
inputs_embeds = self.word_embeddings(input_ids)
token_type_embeddings = self.token_type_embeddings(token_type_ids)
Expand Down Expand Up @@ -525,11 +525,17 @@ def forward(
output_hidden_states=False,
return_dict=True,
):
if self.gradient_checkpointing and self.training and use_cache:
logger.warning(
"`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
)
use_cache = False

all_hidden_states = () if output_hidden_states else None
all_self_attentions = () if output_attentions else None
all_cross_attentions = () if output_attentions and self.config.add_cross_attention else None

next_decoder_cache = () if use_cache else None

for i, layer_module in enumerate(self.layer):
if output_hidden_states:
all_hidden_states = all_hidden_states + (hidden_states,)
Expand All @@ -538,13 +544,6 @@ def forward(
past_key_value = past_key_values[i] if past_key_values is not None else None

if self.gradient_checkpointing and self.training:

if use_cache:
logger.warning(
"`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
)
use_cache = False

def create_custom_forward(module):
def custom_forward(*inputs):
return module(*inputs, past_key_value, output_attentions)
Expand Down Expand Up @@ -2024,7 +2023,7 @@ def _init_weights(self, module):
module.weight.data.normal_(mean=0.0, std=std)
if module.padding_idx is not None:
module.weight.data[module.padding_idx].zero_()

def _set_gradient_checkpointing(self, module, value=False):
if isinstance(module, ({{cookiecutter.camelcase_modelname}}Decoder, {{cookiecutter.camelcase_modelname}}Encoder)):
module.gradient_checkpointing = value
Expand Down Expand Up @@ -2525,6 +2524,10 @@ def forward(
hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training)

# decoder layers
if self.gradient_checkpointing and self.training and use_cache:
logger.warning("`use_cache = True` is incompatible with gradient checkpointing`. Setting `use_cache = False`...")
use_cache = False

all_hidden_states = () if output_hidden_states else None
all_self_attns = () if output_attentions else None
all_cross_attentions = () if (output_attentions and encoder_hidden_states is not None) else None
Expand All @@ -2547,11 +2550,6 @@ def forward(
past_key_value = past_key_values[idx] if past_key_values is not None else None

if self.gradient_checkpointing and self.training:

if use_cache:
logger.warning("`use_cache = True` is incompatible with gradient checkpointing`. Setting `use_cache = False`...")
use_cache = False

def create_custom_forward(module):
def custom_forward(*inputs):
# None for past_key_value
Expand Down