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

Fix circular import in ds_transformer.py #5804

Closed
wants to merge 9 commits into from
Closed

Fix circular import in ds_transformer.py #5804

Changes from 3 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
7f2650c
Fix circular import in ds_transformer.py
sznmelvin Jul 28, 2024
2d766d9
Merge branch 'master' into master
sznmelvin Jul 29, 2024
0be147f
Merge branch 'master' into master
loadams Jul 29, 2024
426445c
Merge branch 'master' into master
sznmelvin Aug 3, 2024
95894bb
update
sznmelvin Aug 3, 2024
dbc10bf
Merge branch 'master' into master
tjruwase Aug 14, 2024
5a6e4b8
Merge branch 'master' into master
loadams Aug 14, 2024
67ce744
Merge branch 'master' into master
loadams Nov 4, 2024
f55e009
Merge branch 'master' into master
loadams Jan 21, 2025
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
10 changes: 7 additions & 3 deletions deepspeed/model_implementations/transformers/ds_transformer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,9 +13,6 @@
from deepspeed.accelerator import get_accelerator
from deepspeed.ops.op_builder import InferenceBuilder
import deepspeed
if deepspeed.HAS_TRITON:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@eonsparks - could you please run the pre-commit formatter (pre-commit run --all-files) to resolve te formatting errors?

from deepspeed.ops.transformer.inference.triton.mlp import TritonMLP
from deepspeed.ops.transformer.inference.triton.attention import TritonSelfAttention

inference_module = None

Expand All @@ -38,6 +35,9 @@ class DeepSpeedTransformerInference(nn.Module):
"""
layer_id = 0

class DeepSpeedTransformerInference(nn.Module):
layer_id = 0

def __init__(self,
config,
mp_group=None,
Expand Down Expand Up @@ -67,12 +67,16 @@ def __init__(self,
assert not self.config.use_triton
else:
if deepspeed.HAS_TRITON and self.config.use_triton:
# Lazy import to avoid circular dependency
from deepspeed.ops.transformer.inference.triton.attention import TritonSelfAttention
self.attention = TritonSelfAttention(self.config)
else:
self.attention = DeepSpeedSelfAttention(self.config, mp_group, quantize_scales, quantize_groups,
merge_count)

if deepspeed.HAS_TRITON and self.config.use_triton:
# Lazy import to avoid circular dependency
from deepspeed.ops.transformer.inference.triton.mlp import TritonMLP
self.mlp = TritonMLP(self.config)
else:
self.mlp = DeepSpeedMLP(self.config, mp_group, quantize_scales, quantize_groups, merge_count,
Expand Down
Loading