Skip to content

Fix lfm2 kernel path #44634

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

Merged
Cyrilvallez merged 2 commits into from
Mar 12, 2026
Merged

Fix lfm2 kernel path #44634

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
f3d6072
fix
Cyrilvallez Mar 12, 2026
ec8f0cc
nit
Cyrilvallez Mar 12, 2026
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
12 changes: 10 additions & 2 deletions src/transformers/models/lfm2/modeling_lfm2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -459,14 +459,22 @@ def cuda_kernels_forward(
past_key_values: Lfm2HybridConvCache | None = None,
attention_mask: torch.Tensor | None = None,
):
seqlen = x.shape[1]
x = apply_mask_to_padding_states(x, attention_mask)
BCx = self.in_proj(x).transpose(-1, -2)
B, C, x = BCx.chunk(3, dim=-2)

Bx = B * x

# Note: we may or may not have to substract the current seq_len here as the cache may or may not be already updated
# by the current layer
past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
# In this case, the cache was already updated and we need to subtract seq_len to get the correct past length
if "full_attention" in self.config.layer_types[: self.layer_idx]:
past_seen_tokens = past_seen_tokens - seqlen

conv_weights = self.conv.weight.view(self.conv.weight.size(0), self.conv.weight.size(2))
if past_key_values is not None and past_key_values.get_seq_length() > 0:
if past_seen_tokens > 0:
conv_out = causal_conv1d_update(
Bx.squeeze(-1),
past_key_values.conv_cache[self.layer_idx],
Expand Down Expand Up @@ -507,7 +515,7 @@ def slow_forward(
if "full_attention" in self.config.layer_types[: self.layer_idx]:
past_seen_tokens = past_seen_tokens - seqlen

if past_key_values is not None and past_seen_tokens > 0:
if past_seen_tokens > 0:
conv_state = past_key_values.conv_cache[self.layer_idx]
cache_position = torch.arange(seqlen, device=conv_state.device) + past_seen_tokens
cache_position = cache_position.clamp(0, self.L_cache - 1)
Expand Down
12 changes: 10 additions & 2 deletions src/transformers/models/lfm2/modular_lfm2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -294,14 +294,22 @@ def cuda_kernels_forward(
past_key_values: Lfm2HybridConvCache | None = None,
attention_mask: torch.Tensor | None = None,
):
seqlen = x.shape[1]
x = apply_mask_to_padding_states(x, attention_mask)
BCx = self.in_proj(x).transpose(-1, -2)
B, C, x = BCx.chunk(3, dim=-2)

Bx = B * x

# Note: we may or may not have to substract the current seq_len here as the cache may or may not be already updated
# by the current layer
past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
# In this case, the cache was already updated and we need to subtract seq_len to get the correct past length
if "full_attention" in self.config.layer_types[: self.layer_idx]:
past_seen_tokens = past_seen_tokens - seqlen

conv_weights = self.conv.weight.view(self.conv.weight.size(0), self.conv.weight.size(2))
if past_key_values is not None and past_key_values.get_seq_length() > 0:
if past_seen_tokens > 0:
conv_out = causal_conv1d_update(
Bx.squeeze(-1),
past_key_values.conv_cache[self.layer_idx],
Expand Down Expand Up @@ -342,7 +350,7 @@ def slow_forward(
if "full_attention" in self.config.layer_types[: self.layer_idx]:
past_seen_tokens = past_seen_tokens - seqlen

if past_key_values is not None and past_seen_tokens > 0:
if past_seen_tokens > 0:
conv_state = past_key_values.conv_cache[self.layer_idx]
cache_position = torch.arange(seqlen, device=conv_state.device) + past_seen_tokens
cache_position = cache_position.clamp(0, self.L_cache - 1)
Expand Down
12 changes: 10 additions & 2 deletions src/transformers/models/lfm2_moe/modeling_lfm2_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -535,14 +535,22 @@ def cuda_kernels_forward(
past_key_values: Lfm2MoeHybridConvCache | None = None,
attention_mask: torch.Tensor | None = None,
):
seqlen = x.shape[1]
x = apply_mask_to_padding_states(x, attention_mask)
BCx = self.in_proj(x).transpose(-1, -2)
B, C, x = BCx.chunk(3, dim=-2)

Bx = B * x

# Note: we may or may not have to substract the current seq_len here as the cache may or may not be already updated
# by the current layer
past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
# In this case, the cache was already updated and we need to subtract seq_len to get the correct past length
if "full_attention" in self.config.layer_types[: self.layer_idx]:
past_seen_tokens = past_seen_tokens - seqlen

conv_weights = self.conv.weight.view(self.conv.weight.size(0), self.conv.weight.size(2))
if past_key_values is not None and past_key_values.get_seq_length() > 0:
if past_seen_tokens > 0:
conv_out = causal_conv1d_update(
Bx.squeeze(-1),
past_key_values.conv_cache[self.layer_idx],
Expand Down Expand Up @@ -583,7 +591,7 @@ def slow_forward(
if "full_attention" in self.config.layer_types[: self.layer_idx]:
past_seen_tokens = past_seen_tokens - seqlen

if past_key_values is not None and past_seen_tokens > 0:
if past_seen_tokens > 0:
conv_state = past_key_values.conv_cache[self.layer_idx]
cache_position = torch.arange(seqlen, device=conv_state.device) + past_seen_tokens
cache_position = cache_position.clamp(0, self.L_cache - 1)
Expand Down
Loading