Skip to content

convert query, key, cos, sin to bf16 if any are bf16 #182

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

Closed
skaulintel wants to merge 2 commits into from
Closed

convert query, key, cos, sin to bf16 if any are bf16 #182

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d597a15
convert query, key, cos, sin to bf16 if any are bf16
skaulintel Apr 30, 2024
0241345
Update modeling_llama.py
skaulintel May 8, 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
31 changes: 19 additions & 12 deletions optimum/habana/transformers/models/llama/modeling_llama.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -981,18 +981,25 @@ def prepare_inputs_for_generation(


def apply_customized_rope(q, k, cos, sin, position_ids, use_fused_rope=True):
if q.device.type == "hpu" and has_fused_rope and use_fused_rope:
# TODO: remove `.clone()` when SynapseAI v1.15 is released
if k.dtype==torch.bfloat16:
return FusedRoPE.apply(
q, cos.unsqueeze(0).unsqueeze(0).clone(), sin.unsqueeze(0).unsqueeze(0).clone(), position_ids
), FusedRoPE.apply(
k, cos.unsqueeze(0).unsqueeze(0).clone().to(torch.bfloat16), sin.unsqueeze(0).unsqueeze(0).clone().to(torch.bfloat16), position_ids
if q.device.type == "hpu" and FusedRoPE:
if q.dtype == torch.bfloat16:
rope_q = FusedRoPE.apply(
q,
cos.unsqueeze(0).unsqueeze(0).to(torch.bfloat16),
sin.unsqueeze(0).unsqueeze(0).to(torch.bfloat16),
position_ids,
)
return FusedRoPE.apply(
q, cos.unsqueeze(0).unsqueeze(0).clone(), sin.unsqueeze(0).unsqueeze(0).clone(), position_ids
), FusedRoPE.apply(
k, cos.unsqueeze(0).unsqueeze(0).clone(), sin.unsqueeze(0).unsqueeze(0).clone(), position_ids
)
else:
rope_q = FusedRoPE.apply(q, cos.unsqueeze(0).unsqueeze(0), sin.unsqueeze(0).unsqueeze(0), position_ids)
if k.dtype == torch.bfloat16:
rope_k = FusedRoPE.apply(
k,
cos.unsqueeze(0).unsqueeze(0).to(torch.bfloat16),
sin.unsqueeze(0).unsqueeze(0).to(torch.bfloat16),
position_ids,
)
else:
rope_k = FusedRoPE.apply(k, cos.unsqueeze(0).unsqueeze(0), sin.unsqueeze(0).unsqueeze(0), position_ids)
return rope_q, rope_k
else:
return apply_rotary_pos_emb(q, k, cos, sin, position_ids)