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

Add aten::_transformer_encoder_layer_fwd operator #1069

Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open

Add aten::_transformer_encoder_layer_fwd operator #1069

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
199b4ce
add _transformer_encoder_layer_fwd
yucai-intel Nov 11, 2024
bed5a45
fix
yucai-intel Nov 12, 2024
a24f44e
Merge branch 'main' into encoder
yucai-intel Nov 12, 2024
55f82c3
skip test cases onednn unsupported
yucai-intel Nov 13, 2024
91a3db7
Merge branch 'main' into encoder
yucai-intel Nov 19, 2024
00d0592
Update skip_list_common.py
yucai-intel Nov 19, 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
1 change: 1 addition & 0 deletions test/xpu/xpu_test_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -219,6 +219,7 @@
"nn.functional.ctc_loss",
"nn.functional.channel_shuffle",
"nn.functional.multi_head_attention_forward",
"nn.TransformerEncoderLayer",
"sigmoid",
"logsigmoid",
"sgn",
Expand Down
7 changes: 7 additions & 0 deletions yaml/native/native_functions.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5859,6 +5859,13 @@
XPU: _dirichlet_grad_xpu
autogen: _dirichlet_grad.out

# Apparently, putting "forward" in the name will cause Python bindings to be skipped, so "fwd" it is.
- func: _transformer_encoder_layer_fwd(Tensor src, int embed_dim, int num_heads, Tensor qkv_weight, Tensor qkv_bias, Tensor proj_weight, Tensor proj_bias, bool use_gelu, bool norm_first, float eps, Tensor norm_weight_1, Tensor norm_bias_1, Tensor norm_weight_2, Tensor norm_bias_2, Tensor ffn_weight_1, Tensor ffn_bias_1, Tensor ffn_weight_2, Tensor ffn_bias_2, Tensor? mask=None, int? mask_type=None) -> Tensor
variants: function
dispatch:
XPU: transformer_encoder_layer_forward
autogen: _transformer_encoder_layer_fwd.out

# Fused implementation detail for transformers. Adds in-projection bias to QKV and divides Q by sqrt(D/num_heads).
- func: _transform_bias_rescale_qkv(Tensor qkv, Tensor qkv_bias, int num_heads) -> (Tensor, Tensor, Tensor)
dispatch:
Expand Down
Loading