Skip to content

fix AutoTP in deepspeed could not work for bloom #22196

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
sgugger merged 2 commits into from
Mar 17, 2023
Merged

fix AutoTP in deepspeed could not work for bloom #22196

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4534821
fix AutoTP in deepspeed could not work for bloom
sywangyi Mar 15, 2023
10af2b4
add a method in BloomModel to build ailib
sywangyi Mar 17, 2023
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
4 changes: 3 additions & 1 deletion src/transformers/models/bloom/modeling_bloom.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -638,6 +638,8 @@ def __init__(self, config: BloomConfig):

self.gradient_checkpointing = False

self.build_alibi_tensor = build_alibi_tensor

# Initialize weights and apply final processing
self.post_init()

Expand Down Expand Up @@ -750,7 +752,7 @@ def forward(
else:
attention_mask = attention_mask.to(hidden_states.device)

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

causal_mask = self._prepare_attn_mask(
attention_mask,
Expand Down