[FIX] Fix Alibi implementation in PagedAttention kernel (#945)

* [FIX] Fix Alibi implementation in PagedAttention kernel

* Fix test_attention

* Fix

---------

Co-authored-by: Woosuk Kwon <[email protected]>
Co-authored-by: Oliver-ss <[email protected]>

Author: 3 people

csrc/attention/attention_kernels.cu

@@ -178,7 +178,7 @@ __global__ void single_query_cached_kv_attention_kernel(
       // This includes a reduction across the threads in the same thread group.
       float qk = scale * Qk_dot<scalar_t, THREAD_GROUP_SIZE>::dot(q_vecs[thread_group_offset], k_vecs);
       // Add the ALiBi bias if slopes are given.
-      qk += (alibi_slope != 0) ? alibi_slope * (token_idx - context_len) : 0;
+      qk += (alibi_slope != 0) ? alibi_slope * (token_idx - context_len + 1) : 0;
 
       if (thread_group_offset == 0) {
         // Store the partial reductions to shared memory.

tests/kernels/test_attention.py

@@ -17,7 +17,7 @@
 NUM_HEADS = [(40, 40), (64, 8)]  # Arbitrary values for testing
 HEAD_SIZES = [64, 80, 96, 112, 128, 256]
 BLOCK_SIZES = [8, 16, 32]
-USE_ALIBI = [False]  # TODO(woosuk): Add USE_ALIBI=True
+USE_ALIBI = [False, True]
 SEEDS = [0]
 
 
@@ -83,7 +83,7 @@ def ref_single_query_cached_kv_attention(
         if alibi_slopes is not None:
             # Create the ALiBi bias used in the paged attention kernel.
             position_ids = torch.arange(context_len, device="cuda").int()
-            alibi_bias = (context_len - position_ids).float()
+            alibi_bias = (position_ids - context_len + 1).float()
             alibi_bias = alibi_slopes.view(-1, 1, 1) * alibi_bias.view(
                 1, 1, -1)
 
@@ -224,6 +224,7 @@ def ref_multi_query_kv_attention(
     return ref_output
 
 
+# TODO(woosuk): Add tests for USE_ALIBI=True.
 @pytest.mark.parametrize("num_seqs", NUM_PREFILL_SEQS)
 @pytest.mark.parametrize("num_heads", NUM_HEADS)
 @pytest.mark.parametrize("head_size", HEAD_SIZES)