[Minimax-M3] Using tok_sparse_select from MSA instead of triton kernels (#47502)
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 4.8
parent
0303f37a54
commit
9021589498
@@ -150,7 +150,7 @@ def _reference_index_topk(
|
||||
num_blocks = (seq_len + BLOCK_SIZE - 1) // BLOCK_SIZE
|
||||
pages = block_table[req_id, :num_blocks]
|
||||
k = index_kv_cache[pages].reshape(num_blocks * BLOCK_SIZE, -1)
|
||||
score = torch.einsum("qhd,kd->hqk", q.float(), k.float()) * sm_scale
|
||||
score = sm_scale * torch.einsum("qhd,kd->hqk", q.float(), k.float())
|
||||
|
||||
q_pos = prefix_len + torch.arange(q_len, device=idx_q.device)
|
||||
k_pos = torch.arange(k.shape[0], device=idx_q.device)
|
||||
@@ -621,11 +621,12 @@ def test_decode_index_topk_fp8(num_idx_heads: int):
|
||||
init_blocks=init_blocks,
|
||||
local_blocks=local_blocks,
|
||||
num_kv_heads=num_idx_heads,
|
||||
sm_scale=head_dim**-0.5,
|
||||
decode_query_len=decode_query_len,
|
||||
max_decode_query_len=decode_query_len,
|
||||
)
|
||||
# Reference from the DEQUANTIZED fp8 values (the kernel computes the fp8 QK
|
||||
# in fp32, so it must match an fp32 matmul of the same e4m3 values).
|
||||
# in fp32 with no scaling, so it must match an unscaled fp32 matmul of the
|
||||
# same e4m3 values).
|
||||
expected = _reference_index_topk(
|
||||
idx_q.float(),
|
||||
index_kv_cache.float(),
|
||||
@@ -636,7 +637,6 @@ def test_decode_index_topk_fp8(num_idx_heads: int):
|
||||
topk,
|
||||
init_blocks,
|
||||
local_blocks,
|
||||
head_dim**-0.5,
|
||||
)
|
||||
_assert_topk_indices_equal_unordered(actual, expected)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user