Skip to content
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

Jump to bottom

[shardformer] update bloom model #5518

Merged
merged 2 commits into from
Apr 1, 2024
Merged

[shardformer] update bloom model #5518

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
1283bf0
update bloom model
wangbluo Mar 27, 2024
0562c97
remove the version restriction
wangbluo Mar 27, 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
39 changes: 15 additions & 24 deletions colossalai/shardformer/modeling/bloom.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@
BloomModel,
)
from transformers.utils import logging

from transformers.modeling_attn_mask_utils import _prepare_4d_causal_attention_mask
from colossalai.pipeline.stage_manager import PipelineStageManager
from colossalai.shardformer.layer._operation import gather_forward_split_backward, split_forward_gather_backward
from colossalai.shardformer.shard import ShardConfig
Expand Down Expand Up @@ -205,12 +205,13 @@ def bloom_model_forward(
alibi = self.build_alibi_tensor(attention_mask, self.num_heads, dtype=hidden_states.dtype)

# causal_mask is constructed every stage and its input is passed through different stages
causal_mask = self._prepare_attn_mask(
causal_mask = _prepare_4d_causal_attention_mask(
attention_mask,
input_shape=(batch_size, seq_length),
inputs_embeds=hidden_states,
past_key_values_length=past_key_values_length,
)

causal_mask = causal_mask.bool()
# split the input tensor along sequence dimension
# [batch_size, seq_len, hidden_size] -> [batch_size, seq_len/TP_size, hidden_size]
if shard_config.enable_sequence_parallelism:
Expand All @@ -226,21 +227,15 @@ def bloom_model_forward(
all_hidden_states = all_hidden_states + (hidden_states,)

if self.gradient_checkpointing and self.training:

def create_custom_forward(module):
def custom_forward(*inputs):
# None for past_key_value
return module(*inputs, use_cache=use_cache, output_attentions=output_attentions)

return custom_forward

outputs = torch.utils.checkpoint.checkpoint(
create_custom_forward(block),
outputs = self._gradient_checkpointing_func(
block.__call__,
hidden_states,
alibi,
causal_mask,
layer_past,
head_mask[i],
use_cache,
output_attentions,
)
else:
outputs = block(
Expand Down Expand Up @@ -1000,11 +995,13 @@ def forward(

alibi = self.build_alibi_tensor(attention_mask, self.num_heads, dtype=hidden_states.dtype)

causal_mask = self._prepare_attn_mask(
causal_mask = _prepare_4d_causal_attention_mask(
attention_mask,
input_shape=(batch_size, seq_length),
inputs_embeds=hidden_states,
past_key_values_length=past_key_values_length,
)
causal_mask = causal_mask.bool()
# split the input tensor along sequence dimension
# [batch_size, seq_len, hidden_size] -> [batch_size, seq_len/TP_size, hidden_size]
hidden_states = split_forward_gather_backward(
Expand All @@ -1016,21 +1013,15 @@ def forward(
all_hidden_states = all_hidden_states + (hidden_states,)

if self.gradient_checkpointing and self.training:

def create_custom_forward(module):
def custom_forward(*inputs):
# None for past_key_value
return module(*inputs, use_cache=use_cache, output_attentions=output_attentions)

return custom_forward

outputs = torch.utils.checkpoint.checkpoint(
create_custom_forward(block),
outputs = self._gradient_checkpointing_func(
block.__call__,
hidden_states,
alibi,
causal_mask,
layer_past,
head_mask[i],
use_cache,
output_attentions,
)
else:
outputs = block(
Expand Down
6 changes: 0 additions & 6 deletions colossalai/shardformer/policies/bloom.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,12 +23,6 @@
class BloomPolicy(Policy):
def __init__(self) -> None:
super().__init__()
import transformers
from packaging.version import Version

assert Version(transformers.__version__) <= Version(
"4.33.0"
), "The Bloom model should run on a transformers version not greater than 4.33.0."

def config_sanity_check(self):
pass
Expand Down
Loading