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 deepseek coder with linear rope type support on GPU #12709

Merged
merged 14 commits into from
Jan 15, 2025
Merged

Fix deepseek coder with linear rope type support on GPU #12709

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
f2fb3df
Fix deepseek coder with linear rope type
Oscilloscope98 Jan 15, 2025
424c761
Style fix
Oscilloscope98 Jan 15, 2025
7106f18
Move to optimize_pre
Oscilloscope98 Jan 15, 2025
cf959c0
Small fix
Oscilloscope98 Jan 15, 2025
15e25e8
Small fix
Oscilloscope98 Jan 15, 2025
b343681
Small fix to not affect other cases
Oscilloscope98 Jan 15, 2025
111d30b
Style fixes
Oscilloscope98 Jan 15, 2025
6c4327e
Update function name
Oscilloscope98 Jan 15, 2025
40cf3fe
Small fix
Oscilloscope98 Jan 15, 2025
3ddfffe
Small fix
Oscilloscope98 Jan 15, 2025
ec344d8
Small fix
Oscilloscope98 Jan 15, 2025
49fcc7b
Fix for low transformers version first
Oscilloscope98 Jan 15, 2025
66de2d3
Style fix
Oscilloscope98 Jan 15, 2025
1ec982a
Small fix
Oscilloscope98 Jan 15, 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
3 changes: 2 additions & 1 deletion python/llm/src/ipex_llm/transformers/convert.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -995,8 +995,9 @@ def _optimize_pre(model, qtype=None):
from ipex_llm.transformers.models.gemma2 import merge_qkv
model.apply(merge_qkv)
elif model.config.model_type == "llama":
from ipex_llm.transformers.models.llama import merge_qkv
from ipex_llm.transformers.models.llama import merge_qkv, pre_compute_inv_freq
model.apply(merge_qkv)
model.apply(pre_compute_inv_freq)
elif model.config.model_type == "mllama":
from ipex_llm.transformers.models.mllama import merge_qkv
model.apply(merge_qkv)
Expand Down
15 changes: 13 additions & 2 deletions python/llm/src/ipex_llm/transformers/models/llama.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -119,6 +119,13 @@ def merge_qkv(module: torch.nn.Module):
merge_qkv_base(module, LlamaAttention)


def pre_compute_inv_freq(module: torch.nn.Module):
if module.__class__.__name__ == "LlamaLinearScalingRotaryEmbedding":
if hasattr(module, "scaling_factor"):
module.register_buffer("inv_freq_scaled", None, persistent=False)
module.inv_freq_scaled = module.inv_freq / module.scaling_factor


def llama_attention_forward(
self,
hidden_states: torch.Tensor,
Expand Down Expand Up @@ -147,8 +154,12 @@ def llama_attention_forward(
import xe_addons
if hasattr(self, "rotary_emb"):
# transformers < 4.46
xe_addons.rotary_half_inplaced(self.rotary_emb.inv_freq, position_ids,
query_states, key_states)
if hasattr(self.rotary_emb, "inv_freq_scaled"):
xe_addons.rotary_half_inplaced(self.rotary_emb.inv_freq_scaled, position_ids,
query_states, key_states)
else:
xe_addons.rotary_half_inplaced(self.rotary_emb.inv_freq, position_ids,
query_states, key_states)
else:
# transformers >= 4.46
cos, sin = position_embeddings
Expand Down
Loading