Skip to content

Fix: num_logits_to_keep in composite models #33168

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
zucchini-nlp merged 2 commits into from
Sep 3, 2024
Merged

Fix: num_logits_to_keep in composite models #33168

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3490a40
fix
zucchini-nlp Aug 28, 2024
97e6a8c
paligemma
zucchini-nlp Aug 28, 2024
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
6 changes: 4 additions & 2 deletions src/transformers/models/cohere/modeling_cohere.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1118,7 +1118,7 @@ def prepare_inputs_for_generation(
cache_position=None,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand Down Expand Up @@ -1169,14 +1169,16 @@ def prepare_inputs_for_generation(
batch_size=batch_size,
)

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
"cache_position": cache_position,
"past_key_values": past_key_values,
"use_cache": use_cache,
"attention_mask": attention_mask,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
6 changes: 4 additions & 2 deletions src/transformers/models/dbrx/modeling_dbrx.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1382,7 +1382,7 @@ def prepare_inputs_for_generation(
cache_position=None,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand Down Expand Up @@ -1433,14 +1433,16 @@ def prepare_inputs_for_generation(
batch_size=batch_size,
)

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
"cache_position": cache_position,
"past_key_values": past_key_values,
"use_cache": use_cache,
"attention_mask": attention_mask,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
6 changes: 4 additions & 2 deletions src/transformers/models/gemma/modeling_gemma.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1131,7 +1131,7 @@ def prepare_inputs_for_generation(
cache_position=None,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand Down Expand Up @@ -1181,14 +1181,16 @@ def prepare_inputs_for_generation(
batch_size=batch_size,
)

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
"cache_position": cache_position,
"past_key_values": past_key_values,
"use_cache": use_cache,
"attention_mask": attention_mask,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
Expand Down
7 changes: 5 additions & 2 deletions src/transformers/models/gemma2/modeling_gemma2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1094,7 +1094,7 @@ def prepare_inputs_for_generation(
cache_position=None,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand Down Expand Up @@ -1148,14 +1148,17 @@ def prepare_inputs_for_generation(
cache_position=cache_position,
batch_size=batch_size,
)

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
"cache_position": cache_position,
"past_key_values": past_key_values,
"use_cache": use_cache,
"attention_mask": attention_mask,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
Expand Down
6 changes: 4 additions & 2 deletions src/transformers/models/idefics2/modeling_idefics2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1643,7 +1643,7 @@ def prepare_inputs_for_generation(
past_key_values=None,
attention_mask=None,
inputs_embeds=None,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
past_length = 0
Expand Down Expand Up @@ -1687,6 +1687,9 @@ def prepare_inputs_for_generation(
else:
model_inputs = {"input_ids": input_ids}

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

image_hidden_states = kwargs.get("image_hidden_states", None)
if image_hidden_states is not None:
pixel_values = None
Expand All @@ -1703,7 +1706,6 @@ def prepare_inputs_for_generation(
"pixel_values": pixel_values,
"pixel_attention_mask": pixel_attention_mask,
"image_hidden_states": image_hidden_states,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
Expand Down
6 changes: 4 additions & 2 deletions src/transformers/models/jetmoe/modeling_jetmoe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1348,7 +1348,7 @@ def prepare_inputs_for_generation(
output_router_logits=False,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand All @@ -1373,6 +1373,9 @@ def prepare_inputs_for_generation(
else:
model_inputs = {"input_ids": input_ids.contiguous()} # `contiguous()` needed for compilation use cases

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
Expand All @@ -1381,7 +1384,6 @@ def prepare_inputs_for_generation(
"use_cache": use_cache,
"attention_mask": attention_mask,
"output_router_logits": output_router_logits,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
Expand Down
6 changes: 4 additions & 2 deletions src/transformers/models/llama/modeling_llama.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1244,7 +1244,7 @@ def prepare_inputs_for_generation(
cache_position=None,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand Down Expand Up @@ -1295,14 +1295,16 @@ def prepare_inputs_for_generation(
batch_size=batch_size,
)

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
"cache_position": cache_position,
"past_key_values": past_key_values,
"use_cache": use_cache,
"attention_mask": attention_mask,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
Expand Down
10 changes: 10 additions & 0 deletions src/transformers/models/llava/modeling_llava.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -377,6 +377,7 @@ def forward(
output_hidden_states: Optional[bool] = None,
return_dict: Optional[bool] = None,
cache_position: Optional[torch.LongTensor] = None,
num_logits_to_keep: int = 0,
) -> Union[Tuple, LlavaCausalLMOutputWithPast]:
r"""
Args:
Expand All @@ -385,6 +386,12 @@ def forward(
config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
(masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.

num_logits_to_keep (`int`, *optional*):
Calculate logits for the last `num_logits_to_keep` tokens. If `0`, calculate logits for all
`input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that
token can save memory, which becomes pretty significant for long sequences or large vocabulary size.


Returns:

Example:
Expand Down Expand Up @@ -518,6 +525,7 @@ def forward(
output_hidden_states=output_hidden_states,
return_dict=return_dict,
cache_position=cache_position,
num_logits_to_keep=num_logits_to_keep,
)

logits = outputs[0]
Expand Down Expand Up @@ -558,6 +566,7 @@ def prepare_inputs_for_generation(
pixel_values=None,
attention_mask=None,
cache_position=None,
num_logits_to_keep=None,
**kwargs,
):
# Trigger the new behavior if we have more than image embeddings seq length tokens for images
Expand All @@ -572,6 +581,7 @@ def prepare_inputs_for_generation(
inputs_embeds=inputs_embeds,
attention_mask=attention_mask,
cache_position=cache_position,
num_logits_to_keep=num_logits_to_keep,
**kwargs,
)

Expand Down
9 changes: 9 additions & 0 deletions src/transformers/models/llava_next/modeling_llava_next.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -721,6 +721,7 @@ def forward(
output_hidden_states: Optional[bool] = None,
return_dict: Optional[bool] = None,
cache_position: Optional[torch.LongTensor] = None,
num_logits_to_keep: int = 0,
) -> Union[Tuple, LlavaNextCausalLMOutputWithPast]:
r"""
Args:
Expand All @@ -729,6 +730,11 @@ def forward(
config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
(masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.

num_logits_to_keep (`int`, *optional*):
Calculate logits for the last `num_logits_to_keep` tokens. If `0`, calculate logits for all
`input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that
token can save memory, which becomes pretty significant for long sequences or large vocabulary size.

Returns:

Example:
Expand Down Expand Up @@ -890,6 +896,7 @@ def forward(
output_hidden_states=output_hidden_states,
return_dict=return_dict,
cache_position=cache_position,
num_logits_to_keep=num_logits_to_keep,
)

logits = outputs[0]
Expand Down Expand Up @@ -931,6 +938,7 @@ def prepare_inputs_for_generation(
image_sizes=None,
attention_mask=None,
cache_position=None,
num_logits_to_keep=None,
**kwargs,
):
legacy_processing = (
Expand All @@ -944,6 +952,7 @@ def prepare_inputs_for_generation(
inputs_embeds=inputs_embeds,
attention_mask=attention_mask,
cache_position=cache_position,
num_logits_to_keep=num_logits_to_keep,
**kwargs,
)

Expand Down
10 changes: 10 additions & 0 deletions src/transformers/models/llava_next_video/modeling_llava_next_video.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -767,6 +767,7 @@ def forward(
output_attentions: Optional[bool] = None,
output_hidden_states: Optional[bool] = None,
return_dict: Optional[bool] = None,
num_logits_to_keep: int = 0,
) -> Union[Tuple, LlavaNextVideoCausalLMOutputWithPast]:
r"""
Args:
Expand All @@ -778,6 +779,10 @@ def forward(
Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
(masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
num_logits_to_keep (`int`, *optional*):
Calculate logits for the last `num_logits_to_keep` tokens. If `0`, calculate logits for all
`input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that
token can save memory, which becomes pretty significant for long sequences or large vocabulary size.

Returns:

Expand Down Expand Up @@ -973,6 +978,7 @@ def forward(
output_attentions=output_attentions,
output_hidden_states=output_hidden_states,
return_dict=return_dict,
num_logits_to_keep=num_logits_to_keep,
)

logits = outputs[0]
Expand Down Expand Up @@ -1014,6 +1020,7 @@ def prepare_inputs_for_generation(
pixel_values_videos=None,
image_sizes=None,
attention_mask=None,
num_logits_to_keep=None,
**kwargs,
):
if past_key_values is not None:
Expand Down Expand Up @@ -1057,6 +1064,9 @@ def prepare_inputs_for_generation(
else:
model_inputs = {"input_ids": input_ids}

if "num_logits_to_keep" != None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
Expand Down
6 changes: 4 additions & 2 deletions src/transformers/models/mistral/modeling_mistral.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1090,7 +1090,7 @@ def prepare_inputs_for_generation(
cache_position=None,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand Down Expand Up @@ -1118,14 +1118,16 @@ def prepare_inputs_for_generation(
else:
model_inputs = {"input_ids": input_ids.contiguous()} # `contiguous()` needed for compilation use cases

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
"cache_position": cache_position,
"past_key_values": past_key_values,
"use_cache": use_cache,
"attention_mask": attention_mask,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
Expand Down
6 changes: 4 additions & 2 deletions src/transformers/models/mixtral/modeling_mixtral.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1348,7 +1348,7 @@ def prepare_inputs_for_generation(
output_router_logits=False,
position_ids=None,
use_cache=True,
num_logits_to_keep=0,
num_logits_to_keep=None,
**kwargs,
):
# If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens
Expand All @@ -1373,6 +1373,9 @@ def prepare_inputs_for_generation(
else:
model_inputs = {"input_ids": input_ids.contiguous()} # `contiguous()` needed for compilation use cases

if num_logits_to_keep is not None:
model_inputs["num_logits_to_keep"] = num_logits_to_keep

model_inputs.update(
{
"position_ids": position_ids,
Expand All @@ -1381,7 +1384,6 @@ def prepare_inputs_for_generation(
"use_cache": use_cache,
"attention_mask": attention_mask,
"output_router_logits": output_router_logits,
"num_logits_to_keep": num_logits_to_keep,
}
)
return model_inputs
Expand Down
Loading