[shardformer] hotfix attn mask (#5947)

hpcaitech · Jul 29, 2024 · 7b38964 · 7b38964
1 parent 9664b1b
commit 7b38964
Showing 1 changed file with 7 additions and 7 deletions.
diff --git a/colossalai/shardformer/layer/attn.py b/colossalai/shardformer/layer/attn.py
@@ -139,12 +139,11 @@ def prepare_attn_kwargs(
             # no padding
             assert is_causal
             outputs["attention_mask_type"] = AttnMaskType.CAUSAL
-            attention_mask = torch.ones(s_q, s_kv, dtype=dtype, device=device).tril(diagonal=0).expand(b, s_q, s_kv)
+            attention_mask = torch.ones(s_q, s_kv, dtype=dtype, device=device)
+            if s_q != 1:
+                attention_mask = attention_mask.tril(diagonal=0)
+            attention_mask = attention_mask.expand(b, s_q, s_kv)
         else:
-            assert q_padding_mask.shape == (
-                b,
-                s_q,
-            ), f"q_padding_mask shape {q_padding_mask.shape} should be the same. ({shape_4d})"
             max_seqlen_q, cu_seqlens_q, q_indices = get_pad_info(q_padding_mask)
             if kv_padding_mask is None:
                 # self attention
@@ -156,7 +155,7 @@ def prepare_attn_kwargs(
                 b,
                 s_kv,
             ), f"q_padding_mask shape {kv_padding_mask.shape} should be the same. ({shape_4d})"
-            attention_mask = q_padding_mask[:, None, :].expand(b, s_kv, s_q).to(dtype=dtype, device=device)
+            attention_mask = kv_padding_mask[:, None, :].expand(b, s_q, s_kv).to(dtype=dtype, device=device)
             outputs.update(
                 {
                     "cu_seqlens_q": cu_seqlens_q,
@@ -169,7 +168,8 @@ def prepare_attn_kwargs(
             )
             if is_causal:
                 outputs["attention_mask_type"] = AttnMaskType.PADDED_CAUSAL
-                attention_mask = attention_mask * attention_mask.new_ones(s_q, s_kv).tril(diagonal=0)
+                if s_q != 1:
+                    attention_mask = attention_mask * attention_mask.new_ones(s_q, s_kv).tril(diagonal=0)
             else:
                 outputs["attention_mask_type"] = AttnMaskType.PADDED
         attention_mask = invert_mask(attention_mask).unsqueeze(1)