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

GemmaMLP uses 'tanh` approximation for GeLU activation #1004

Merged
merged 2 commits into from
Mar 5, 2024
Merged

GemmaMLP uses 'tanh` approximation for GeLU activation #1004

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8dac6cc
GemmaMLP uses 'tanh` approximation for GeLU activation
Andrei-Aksionov Mar 5, 2024
b675de2
Force HF in tests to use `tanh` approximation
Andrei-Aksionov Mar 5, 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
2 changes: 2 additions & 0 deletions lit_gpt/config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -866,6 +866,7 @@ def norm_class(self) -> Type:
bias=False,
_norm_class="RMSNorm",
_mlp_class="GemmaMLP",
gelu_approximate="tanh",
intermediate_size=16384,
),
# https://huggingface.co/google/gemma-7b/blob/main/config.json
Expand All @@ -884,6 +885,7 @@ def norm_class(self) -> Type:
bias=False,
_norm_class="RMSNorm",
_mlp_class="GemmaMLP",
gelu_approximate="tanh",
intermediate_size=24576,
),
]
Expand Down
4 changes: 3 additions & 1 deletion lit_gpt/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -287,6 +287,8 @@ def __init__(self, config: Config) -> None:
self.fc_2 = nn.Linear(config.n_embd, config.intermediate_size, bias=config.bias)
self.proj = nn.Linear(config.intermediate_size, config.n_embd, bias=config.bias)

self.config = config

def forward(self, x: torch.Tensor) -> torch.Tensor:
x_fc_1 = self.fc_1(x)
x_fc_2 = self.fc_2(x)
Expand All @@ -298,7 +300,7 @@ class GemmaMLP(LLaMAMLP):
def forward(self, x: torch.Tensor) -> torch.Tensor:
x_fc_1 = self.fc_1(x)
x_fc_2 = self.fc_2(x)
x = torch.nn.functional.gelu(x_fc_1) * x_fc_2
x = torch.nn.functional.gelu(x_fc_1, approximate=self.config.gelu_approximate) * x_fc_2
return self.proj(x)


Expand Down
1 change: 1 addition & 0 deletions tests/test_convert_lit_checkpoint.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -414,6 +414,7 @@ def test_against_original_gemma(model_name, device, dtype):
rope_theta=ours_config.rope_base,
attention_bias=ours_config.bias,
tie_word_embeddings=True,
hidden_act="gelu_pytorch_tanh",
)
assert ours_config.intermediate_size == theirs_config.intermediate_size

Expand Down
1 change: 1 addition & 0 deletions tests/test_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -596,6 +596,7 @@ def test_against_original_gemma(model_name, device, dtype):
rope_theta=ours_config.rope_base,
attention_bias=ours_config.bias,
tie_word_embeddings=True,
hidden_act="gelu_pytorch_tanh",
carmocca marked this conversation as resolved.
Show resolved Hide resolved
)
assert ours_config.intermediate_size == theirs_config.intermediate_size

Expand Down
Loading