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.

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

Remove smaller limit for legacy bfloat16 serialization #505

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from
Draft
Changes from all commits
Commits
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
3 changes: 1 addition & 2 deletions src/petals/client/remote_forward_backward.py
Original file line number Diff line number Diff line change
Expand Up @@ -104,8 +104,7 @@ async def run_remote_forward(

# call RPC on remote server
size = sum(t.element_size() * t.nelement() for t in inputs)
forward_fn = _forward_stream if size > MAX_UNARY_PAYLOAD_SIZE // 2 else _forward_unary
# Hotfix: we use "// 2" since hivemind==1.1.5 serializes bfloat16 tensors in float32, so they take 2x more space
forward_fn = _forward_stream if size > MAX_UNARY_PAYLOAD_SIZE else _forward_unary
deserialized_outputs = await forward_fn(uid, serialized_tensors, stub, config, metadata=metadata, **kwargs)
return nested_pack(deserialized_outputs, structure=rpc_info["outputs_schema"])

Expand Down