ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

[bugs] fix for casual mask (#2868)

Browse source 
* fix for casual mask

* use un_casual in sdpa

* add missing mask

* fix for type
This commit is contained in:
Lei Zhenyuan 2025-07-10 05:10:25 +08:00 • committed by GitHub
commit 78a5e58932
1 changed files with 16 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

18
unsloth/models/llama.py
View file

@ -324,6 +324,13 @@ def LlamaAttention_fast_forward_inference(
# Knn, Vnn = Knn, Vnn
# pass
# when qlen==vlen and attn_mask is None, we should use causal attention
Q_len = Qn.shape[-2]
K_len = Knn.shape[-2]
if attention_mask is None and Q_len == K_len:
is_causal = True
else:
is_causal = False
# Attention
if bsz == 1:
Qn *= self.scalar # See https://github.com/ggerganov/llama.cpp/issues/7805#issuecomment-2153349963
@ -524,11 +531,18 @@ def LlamaAttention_fast_forward(
V = V.transpose(1, 2)
A = flash_attn_func(Q, K, V, causal = True)
else:
# when qlen==vlen and attn_mask is None, we should use causal attention
Q_len = Q.shape[-2]
K_len = K.shape[-2]
if attention_mask is None and Q_len == K_len:
is_causal = True
else:
is_causal = False
# Grouped query attention
if SDPA_HAS_GQA:
# Needs (batch_size, n_heads, seq_len, head_dim)
# is_casual and attention_mask must not be both set!
A = scaled_dot_product_attention(Q, K, V, attn_mask = attention_mask, is_causal = False, enable_gqa = n_groups != 1)
A = scaled_dot_product_attention(Q, K, V, attn_mask = attention_mask, is_causal = is_causal, enable_gqa = n_groups != 1)
# Go back to (batch_size, seq_len, n_heads, head_dim)
A = A.transpose(1, 2)#.contiguous()
else:
@ -543,7 +557,7 @@ def LlamaAttention_fast_forward(
Q, K, V = Q.contiguous(), K.contiguous(), V.contiguous()
# Needs (batch_size, n_heads, seq_len, head_dim)
# is_casual and attention_mask must not be both set!
A = scaled_dot_product_attention(Q, K, V, attn_mask = attention_mask, is_causal = False)
A = scaled_dot_product_attention(Q, K, V, attn_mask = attention_mask, is_causal = is_causal)
# Go back to (batch_size, seq_len, n_heads, head_dim)
A = A.transpose(1, 2).contiguous()
pass