Skip to content

GitLab

"git@developer.sourcefind.cn:zhaoyu6/sglang.git" did not exist on "419a57e771197b14dbe1b5d9b8eacec4ab517d5b"
Commit 5ab9b366 authored by Tri Dao's avatar Tri Dao
Browse files

Clean up alibi, implement non-causal alibi

parent bc28eacc
Hide whitespace changes
Inline Side-by-side
Showing with 309 additions and 1199 deletions
csrc/flash_attn/flash_api.cpp
View file @ 5ab9b366
......@@ -253,12 +253,12 @@ mha_fwd(at::Tensor &q, // batch_size x seqlen_q x num_heads x head_size
const at::Tensor &k, // batch_size x seqlen_k x num_heads_k x head_size
const at::Tensor &v, // batch_size x seqlen_k x num_heads_k x head_size
c10::optional<at::Tensor> &out_, // batch_size x seqlen_q x num_heads x head_size
c10::optional<at::Tensor> &alibi_slopes_, // num_heads or batch_size x num_heads
const float p_dropout,
const float softmax_scale,
bool is_causal,
const int window_size_left,
int window_size_right,
c10::optional<at::Tensor> &alibi_slopes_, // batch_size x num_heads
const bool return_softmax,
c10::optional<at::Generator> gen_) {
......@@ -297,13 +297,13 @@ mha_fwd(at::Tensor &q, // batch_size x seqlen_q x num_heads x head_size
TORCH_CHECK(head_size_og <= 256, "FlashAttention forward only supports head dimension at most 256");
TORCH_CHECK(num_heads % num_heads_k == 0, "Number of heads in key/value must divide number of heads in query");
if (seqlen_q == 1) { is_causal = false; } // causal=true is the same as causal=false in this case
// causal=true is the same as causal=false in this case
if (seqlen_q == 1 && !alibi_slopes_.has_value()) { is_causal = false; }
if (is_causal) { window_size_right = 0; }
// Faster to transpose q from (b, 1, (nheads_kv ngroups), d) to (b, ngroups, nheads_kv, d) in this case
// H/t Daniel Haziza
// TODO: how to make "seqlenq_ngroups_swapped" and ALiBi work together?
const int seqlenq_ngroups_swapped = seqlen_q == 1 && num_heads > num_heads_k && window_size_left < 0 && window_size_right < 0 && p_dropout == 0.f && head_size_og % 8 == 0 && !(alibi_slopes_.has_value());
const int seqlenq_ngroups_swapped = seqlen_q == 1 && num_heads > num_heads_k && window_size_left < 0 && window_size_right < 0 && p_dropout == 0.f && head_size_og % 8 == 0 && !alibi_slopes_.has_value();
if (seqlenq_ngroups_swapped) {
const int ngroups = num_heads / num_heads_k;
q = q.reshape({batch_size, num_heads_k, ngroups, head_size_og}).transpose(1, 2);
......@@ -416,12 +416,11 @@ mha_fwd(at::Tensor &q, // batch_size x seqlen_q x num_heads x head_size
TORCH_CHECK(alibi_slopes.dtype() == torch::kFloat32, "ALiBi slopes must have dtype fp32");
CHECK_DEVICE(alibi_slopes);
TORCH_CHECK(alibi_slopes.stride(-1) == 1, "ALiBi slopes tensor must have contiguous last dimension");
CHECK_SHAPE(alibi_slopes, batch_size, num_heads);
params.has_alibi = true;
TORCH_CHECK(alibi_slopes.sizes() == torch::IntArrayRef({num_heads}) || alibi_slopes.sizes() == torch::IntArrayRef({batch_size, num_heads}));
params.alibi_slopes_ptr = alibi_slopes.data_ptr();
params.alibi_slopes_batch_stride = alibi_slopes.stride(0);
params.alibi_slopes_batch_stride = alibi_slopes.dim() == 2 ? alibi_slopes.stride(0) : 0;
} else {
params.has_alibi = false;
params.alibi_slopes_ptr = nullptr;
}
if (seqlen_k > 0) {
......@@ -456,6 +455,7 @@ mha_varlen_fwd(const at::Tensor &q, // total_q x num_heads x head_size, total_q
const at::Tensor &cu_seqlens_q, // b+1
const at::Tensor &cu_seqlens_k, // b+1
c10::optional<at::Tensor> &seqused_k, // b. If given, only this many elements of each batch element's keys are used.
c10::optional<at::Tensor> &alibi_slopes_, // num_heads or b x num_heads
const int max_seqlen_q,
const int max_seqlen_k,
const float p_dropout,
......@@ -464,7 +464,6 @@ mha_varlen_fwd(const at::Tensor &q, // total_q x num_heads x head_size, total_q
const bool is_causal,
const int window_size_left,
int window_size_right,
c10::optional<at::Tensor> &alibi_slopes_, // b x num_heads
const bool return_softmax,
c10::optional<at::Generator> gen_) {
......@@ -612,12 +611,11 @@ mha_varlen_fwd(const at::Tensor &q, // total_q x num_heads x head_size, total_q
TORCH_CHECK(alibi_slopes.dtype() == torch::kFloat32, "ALiBi slopes must have dtype fp32");
CHECK_DEVICE(alibi_slopes);
TORCH_CHECK(alibi_slopes.stride(-1) == 1, "ALiBi slopes tensor must have contiguous last dimension");
CHECK_SHAPE(alibi_slopes, batch_size, num_heads);
params.has_alibi = true;
TORCH_CHECK(alibi_slopes.sizes() == torch::IntArrayRef({num_heads}) || alibi_slopes.sizes() == torch::IntArrayRef({batch_size, num_heads}));
params.alibi_slopes_ptr = alibi_slopes.data_ptr();
params.alibi_slopes_batch_stride = alibi_slopes.stride(0);
params.alibi_slopes_batch_stride = alibi_slopes.dim() == 2 ? alibi_slopes.stride(0) : 0;
} else {
params.has_alibi = false;
params.alibi_slopes_ptr = nullptr;
}
auto stream = at::cuda::getCurrentCUDAStream().stream();
......@@ -664,12 +662,12 @@ mha_bwd(const at::Tensor &dout, // batch_size x seqlen_q x num_heads, x head_si
c10::optional<at::Tensor> &dq_, // batch_size x seqlen_q x num_heads x head_size
c10::optional<at::Tensor> &dk_, // batch_size x seqlen_k x num_heads_k x head_size
c10::optional<at::Tensor> &dv_, // batch_size x seqlen_k x num_heads_k x head_size
c10::optional<at::Tensor> &alibi_slopes_, // num_heads or batch_size x num_heads
const float p_dropout, // probability to drop
const float softmax_scale,
const bool is_causal,
const int window_size_left,
int window_size_right,
c10::optional<at::Tensor> &alibi_slopes_, // batch_size x num_heads
c10::optional<at::Generator> gen_,
c10::optional<at::Tensor> &rng_state) {
......@@ -848,12 +846,11 @@ mha_bwd(const at::Tensor &dout, // batch_size x seqlen_q x num_heads, x head_si
TORCH_CHECK(alibi_slopes.dtype() == torch::kFloat32, "ALiBi slopes must have dtype fp32");
CHECK_DEVICE(alibi_slopes);
TORCH_CHECK(alibi_slopes.stride(-1) == 1, "ALiBi slopes tensor must have contiguous last dimension");
CHECK_SHAPE(alibi_slopes, batch_size, num_heads);
params.has_alibi = true;
TORCH_CHECK(alibi_slopes.sizes() == torch::IntArrayRef({num_heads}) || alibi_slopes.sizes() == torch::IntArrayRef({batch_size, num_heads}));
params.alibi_slopes_ptr = alibi_slopes.data_ptr();
params.alibi_slopes_batch_stride = alibi_slopes.stride(0);
params.alibi_slopes_batch_stride = alibi_slopes.dim() == 2 ? alibi_slopes.stride(0) : 0;
} else {
params.has_alibi = false;
params.alibi_slopes_ptr = nullptr;
}
if (seqlen_q > 0) {
......@@ -891,6 +888,7 @@ mha_varlen_bwd(const at::Tensor &dout, // total_q x num_heads, x head_size
c10::optional<at::Tensor> &dv_, // total_k x num_heads_k x head_size, total_k := \sum_{i=0}^{b} s_i
const at::Tensor &cu_seqlens_q, // b+1
const at::Tensor &cu_seqlens_k, // b+1
c10::optional<at::Tensor> &alibi_slopes_, // num_heads or b x num_heads
const int max_seqlen_q,
const int max_seqlen_k, // max sequence length to choose the kernel
const float p_dropout, // probability to drop
......@@ -899,7 +897,6 @@ mha_varlen_bwd(const at::Tensor &dout, // total_q x num_heads, x head_size
const bool is_causal,
const int window_size_left,
int window_size_right,
c10::optional<at::Tensor> &alibi_slopes_, // b x num_heads
c10::optional<at::Generator> gen_,
c10::optional<at::Tensor> &rng_state) {
......@@ -1094,12 +1091,11 @@ mha_varlen_bwd(const at::Tensor &dout, // total_q x num_heads, x head_size
TORCH_CHECK(alibi_slopes.dtype() == torch::kFloat32, "ALiBi slopes must have dtype fp32");
CHECK_DEVICE(alibi_slopes);
TORCH_CHECK(alibi_slopes.stride(-1) == 1, "ALiBi slopes tensor must have contiguous last dimension");
CHECK_SHAPE(alibi_slopes, batch_size, num_heads);
params.has_alibi = true;
TORCH_CHECK(alibi_slopes.sizes() == torch::IntArrayRef({num_heads}) || alibi_slopes.sizes() == torch::IntArrayRef({batch_size, num_heads}));
params.alibi_slopes_ptr = alibi_slopes.data_ptr();
params.alibi_slopes_batch_stride = alibi_slopes.stride(0);
params.alibi_slopes_batch_stride = alibi_slopes.dim() == 2 ? alibi_slopes.stride(0) : 0;
} else {
params.has_alibi = false;
params.alibi_slopes_ptr = nullptr;
}
launch(params, stream, /*configure=*/false);
......@@ -1128,14 +1124,14 @@ mha_fwd_kvcache(at::Tensor &q, // batch_size x seqlen_q x num_he
c10::optional<const at::Tensor> &rotary_cos_, // seqlen_ro x (rotary_dim / 2)
c10::optional<const at::Tensor> &rotary_sin_, // seqlen_ro x (rotary_dim / 2)
c10::optional<const at::Tensor> &cache_batch_idx_, // indices to index into the KV cache
c10::optional<at::Tensor> &alibi_slopes_, // num_heads or batch_size x num_heads
c10::optional<at::Tensor> &out_, // batch_size x seqlen_q x num_heads x head_size
const float softmax_scale,
bool is_causal,
const int window_size_left,
int window_size_right,
bool is_rotary_interleaved, // if true, rotary combines indices 0 & 1, else indices 0 & rotary_dim / 2
int num_splits,
c10::optional<at::Tensor> &alibi_slopes_ // batch_size x num_heads
int num_splits
) {
auto dprops = at::cuda::getCurrentDeviceProperties();
......@@ -1174,13 +1170,13 @@ mha_fwd_kvcache(at::Tensor &q, // batch_size x seqlen_q x num_he
TORCH_CHECK(head_size_og <= 256, "FlashAttention forward only supports head dimension at most 256");
TORCH_CHECK(num_heads % num_heads_k == 0, "Number of heads in key/value must divide number of heads in query");
if (seqlen_q == 1) { is_causal = false; } // causal=true is the same as causal=false in this case
// causal=true is the same as causal=false in this case
if (seqlen_q == 1 && !alibi_slopes_.has_value()) { is_causal = false; }
if (is_causal) { window_size_right = 0; }
// Faster to transpose q from (b, 1, (nheads_kv ngroups), d) to (b, ngroups, nheads_kv, d) in this case
// H/t Daniel Haziza
// TODO: how to make "seqlenq_ngroups_swapped" and ALiBi work together?
const int seqlenq_ngroups_swapped = seqlen_q == 1 && num_heads > num_heads_k && window_size_left < 0 && window_size_right < 0 && head_size_og % 8 == 0 && !(alibi_slopes_.has_value());
const int seqlenq_ngroups_swapped = seqlen_q == 1 && num_heads > num_heads_k && window_size_left < 0 && window_size_right < 0 && head_size_og % 8 == 0 && !alibi_slopes_.has_value();
if (seqlenq_ngroups_swapped) {
const int ngroups = num_heads / num_heads_k;
q = q.reshape({batch_size, num_heads_k, ngroups, head_size_og}).transpose(1, 2);
......@@ -1347,12 +1343,11 @@ mha_fwd_kvcache(at::Tensor &q, // batch_size x seqlen_q x num_he
TORCH_CHECK(alibi_slopes.dtype() == torch::kFloat32, "ALiBi slopes must have dtype fp32");
CHECK_DEVICE(alibi_slopes);
TORCH_CHECK(alibi_slopes.stride(-1) == 1, "ALiBi slopes tensor must have contiguous last dimension");
CHECK_SHAPE(alibi_slopes, batch_size, num_heads);
params.has_alibi = true;
TORCH_CHECK(alibi_slopes.sizes() == torch::IntArrayRef({num_heads}) || alibi_slopes.sizes() == torch::IntArrayRef({batch_size, num_heads}));
params.alibi_slopes_ptr = alibi_slopes.data_ptr();
params.alibi_slopes_batch_stride = alibi_slopes.stride(0);
params.alibi_slopes_batch_stride = alibi_slopes.dim() == 2 ? alibi_slopes.stride(0) : 0;
} else {
params.has_alibi = false;
params.alibi_slopes_ptr = nullptr;
}
auto stream = at::cuda::getCurrentCUDAStream().stream();
......
csrc/flash_attn/src/alibi.h
View file @ 5ab9b366
......@@ -13,37 +13,45 @@ using namespace cute;
////////////////////////////////////////////////////////////////////////////////////////////////////
template <typename Engine, typename Layout>
template <bool Is_causal, typename Engine, typename Layout>
inline __device__ void apply_alibi(Tensor<Engine, Layout> &tensor,
const int col_idx_offset_,
const int max_seqlen_k,
const int row_idx_offset_,
const int row_idx_offset,
const int max_seqlen_q,
const int warp_row_stride,
const int head_idx,
const float softmax_scale,
const float alibi_slope) {
// tensor has shape (ncol=(2, MMA_M), nrow=(2, MMA_N))
static_assert(Layout::rank == 2, "Only support 2D Tensor");
const int lane_id = threadIdx.x % 32;
const int row_idx_offset = row_idx_offset_;
const int col_idx_offset = col_idx_offset_ + (lane_id % 4) * 2;
const float alibi_slope_unscaled = alibi_slope / softmax_scale;
#pragma unroll
for (int mi = 0; mi < size<0, 1>(tensor); ++mi) {
const int row_idx_base = row_idx_offset + mi * warp_row_stride;
if constexpr (Is_causal) { // Simpler, we add the same bias vector to all rows
#pragma unroll
for (int i = 0; i < size<0, 0>(tensor); ++i) {
const int row_idx = row_idx_base + i * 8;
for (int nj = 0; nj < size<1, 1>(tensor); ++nj) {
const int col_idx_base = col_idx_offset + nj * 8;
#pragma unroll
for (int nj = 0; nj < size<1, 1>(tensor); ++nj) {
const int col_idx_base = col_idx_offset + nj * 8;
for (int j = 0; j < size<1, 0>(tensor); ++j) {
const int col_idx = col_idx_base + j;
#pragma unroll
for (int j = 0; j < size<1, 0>(tensor); ++j) {
const int col_idx = col_idx_base + j;
const float alibi = alibi_slope_unscaled * col_idx;
if (col_idx < max_seqlen_k && row_idx < max_seqlen_q) {
tensor(make_coord(i, mi), make_coord(j, nj)) += alibi;
for (int mi = 0; mi < size<0>(tensor); ++mi) {
tensor(mi, make_coord(j, nj)) += alibi_slope * col_idx;
}
}
}
} else { // Bias depends on both row_idx and col_idx
#pragma unroll
for (int mi = 0; mi < size<0, 1>(tensor); ++mi) {
const int row_idx_base = row_idx_offset + mi * warp_row_stride;
#pragma unroll
for (int i = 0; i < size<0, 0>(tensor); ++i) {
const int row_idx = row_idx_base + i * 8;
#pragma unroll
for (int nj = 0; nj < size<1, 1>(tensor); ++nj) {
const int col_idx_base = col_idx_offset + nj * 8;
#pragma unroll
for (int j = 0; j < size<1, 0>(tensor); ++j) {
const int col_idx = col_idx_base + j;
tensor(make_coord(i, mi), make_coord(j, nj)) -= alibi_slope * abs(row_idx + max_seqlen_k - max_seqlen_q - col_idx);
}
}
}
......@@ -51,4 +59,4 @@ inline __device__ void apply_alibi(Tensor<Engine, Layout> &tensor,
}
}
} // namespace flash
\ No newline at end of file
} // namespace flash
csrc/flash_attn/src/flash.h
View file @ 5ab9b366
......@@ -131,10 +131,6 @@ struct Flash_fwd_params : public Qkv_params {
int num_splits; // For split-KV version
// float alibi_start;
// float alibi_ratio;
bool has_alibi;
void * __restrict__ alibi_slopes_ptr;
index_t alibi_slopes_batch_stride;
};
......
csrc/flash_attn/src/flash_bwd_kernel.h
View file @ 5ab9b366
......@@ -753,8 +753,12 @@ inline __device__ void compute_dq_dk_dv_1colblock(const Params &params, const in
#pragma unroll
for (int mi = 0; mi < size(lse); ++mi) {
const int row = get<0>(taccScS_row(mi));
lse(mi) = Is_even_MN || row < binfo.actual_seqlen_q - m_block * kBlockM ? gLSE(row) : 0;
lse(mi) = Is_even_MN || row < binfo.actual_seqlen_q - m_block * kBlockM ? gLSE(row) : INFINITY;
}
// We want LSE = inf if the row is OOB. In that case Q would be zero, K would be zero,
// and scores would be zero. With LSE = 0, probs will be all 1's, and when we multiply
// with V (which would be zero), we're fine. However, with ALiBi, we might modify these
// scores, and probs can become NaN. Instead if we set LSE = inf for OOB rows, probs are always 0.
// Tensor tKrK = make_fragment_like(tKsK);
// // cute::copy(gmem_tiled_copy_QKV, tKgK(_, _, _, 0), tKrK);
......@@ -792,18 +796,7 @@ inline __device__ void compute_dq_dk_dv_1colblock(const Params &params, const in
clear(acc_dv);
clear(acc_dk);
float alibi_slope = 0.0f;
if (Has_alibi) {
Tensor gAS = make_tensor(
make_gmem_ptr(
reinterpret_cast<ElementAccum *>(params.alibi_slopes_ptr)
+ bidb * params.alibi_slopes_batch_stride + bidh
),
Shape<_1>{});
Tensor rAS = make_fragment_like(gAS);
cute::copy(gAS, rAS);
alibi_slope = rAS(0);
}
float alibi_slope = !Has_alibi ? 0.0f : reinterpret_cast<float *>(params.alibi_slopes_ptr)[bidb * params.alibi_slopes_batch_stride + bidh] / params.scale_softmax;
for (; m_block >= m_block_min; --m_block) {
Tensor acc_s = partition_fragment_C(tiled_mma_sdp, Shape<Int<kBlockM>, Int<kBlockN>>{}); // (MMA=4, MMA_N, MMA_N)
......@@ -830,18 +823,17 @@ inline __device__ void compute_dq_dk_dv_1colblock(const Params &params, const in
// if (cute::thread(32, 0)) { print(scores); }
if (Has_alibi) {
flash::apply_alibi(
flash::apply_alibi<Is_causal>(
scores,
n_block * kBlockN + (tidx / 32 / AtomLayoutMS) * MMA_N_SdP * 16,
binfo.actual_seqlen_k,
m_block * kBlockM + get<0>(taccScS_row(0)),
binfo.actual_seqlen_q,
AtomLayoutMS * 16,
bidh, params.scale_softmax,
alibi_slope
);
}
// TD [2023-07-29]: I was thinking that we don't need to mask out the elements beyond
// actual_seqlen_k, because acc_s would be some finite value for those indices.
// In the end when we multiply with K to get dQ, the corresponding values of K would be 0,
......@@ -1403,18 +1395,7 @@ inline __device__ void compute_dq_dk_dv_1rowblock(const Params &params, const in
clear(acc_dq);
float alibi_slope = 0.0f;
if (Has_alibi) {
Tensor gAS = make_tensor(
make_gmem_ptr(
reinterpret_cast<ElementAccum *>(params.alibi_slopes_ptr)
+ bidb * params.alibi_slopes_batch_stride + bidh
),
Shape<_1>{});
Tensor rAS = make_fragment_like(gAS);
cute::copy(gAS, rAS);
alibi_slope = rAS(0);
}
float alibi_slope = !Has_alibi ? 0.0f : reinterpret_cast<float *>(params.alibi_slopes_ptr)[bidb * params.alibi_slopes_batch_stride + bidh] / params.scale_softmax;
for (; n_block >= 0; --n_block) {
Tensor acc_s = partition_fragment_C(tiled_mma_sdp, Shape<Int<kBlockM>, Int<kBlockN>>{}); // (MMA=4, MMA_M_SdP, MMA_N)
......@@ -1429,14 +1410,13 @@ inline __device__ void compute_dq_dk_dv_1rowblock(const Params &params, const in
Tensor scores = make_tensor(acc_s.data(), flash::convert_layout_acc_rowcol(acc_s.layout()));
if (Has_alibi) {
flash::apply_alibi(
flash::apply_alibi<Is_causal>(
scores,
n_block * kBlockN + (tidx / 32 / AtomLayoutMS) * MMA_N_SdP * 16,
binfo.actual_seqlen_k,
m_block * kBlockM + get<0>(taccScS_row(0)),
binfo.actual_seqlen_q,
AtomLayoutMS * 16,
bidh, params.scale_softmax,
alibi_slope
);
}
......
csrc/flash_attn/src/flash_bwd_launch_template.h
View file @ 5ab9b366
......@@ -64,12 +64,11 @@ void run_flash_bwd_seqk_parallel(Flash_bwd_params &params, cudaStream_t stream,
BOOL_SWITCH(is_even_MN, IsEvenMNConst, [&] {
BOOL_SWITCH(is_even_K, IsEvenKConst, [&] {
BOOL_SWITCH((params.window_size_left >= 0 || params.window_size_right >= 0) && !params.is_causal, Is_local, [&] {
BOOL_SWITCH(params.has_alibi, Has_alibi, [&] {
BOOL_SWITCH(params.alibi_slopes_ptr != nullptr, Has_alibi, [&] {
// If not IsEvenKConst, we also set IsEvenMNConst to false to reduce number of templates.
// If head dim > 128, set IsEvenMNConst to false to reduce number of templates
// If Is_local, set Is_causal to false
auto kernel = &flash_bwd_dq_dk_dv_loop_seqk_parallel_kernel<Kernel_traits, Is_dropout, Is_causal && !Is_local, Is_local, Has_alibi, IsEvenMNConst && IsEvenKConst && !Is_local && Kernel_traits::kHeadDim <= 128, IsEvenKConst>;
// auto kernel = &flash_bwd_dq_dk_dv_loop_seqk_parallel_kernel<Kernel_traits, Is_dropout, IsCausalConst, IsEvenMNConst, true>;
auto kernel = &flash_bwd_dq_dk_dv_loop_seqk_parallel_kernel<Kernel_traits, Is_dropout, Is_causal, Is_local && !Is_causal, Has_alibi, IsEvenMNConst && IsEvenKConst && !Is_local && Kernel_traits::kHeadDim <= 128, IsEvenKConst>;
if (smem_size_dq_dk_dv >= 48 * 1024) {
C10_CUDA_CHECK(cudaFuncSetAttribute(
kernel, cudaFuncAttributeMaxDynamicSharedMemorySize, smem_size_dq_dk_dv));
......@@ -109,7 +108,7 @@ void run_flash_bwd_seqq_parallel(Flash_bwd_params &params, cudaStream_t stream,
BOOL_SWITCH(params.is_causal, Is_causal, [&] {
BOOL_SWITCH(is_even_N, IsEvenNConst, [&] {
BOOL_SWITCH(is_even_K, IsEvenKConst, [&] {
BOOL_SWITCH(params.has_alibi, Has_alibi, [&] {
BOOL_SWITCH(params.alibi_slopes_ptr != nullptr, Has_alibi, [&] {
// If not IsEvenKConst, we also set IsEvenMNConst to false to reduce number of templates.
auto kernel = &flash_bwd_dq_dk_dv_loop_seqq_parallel_kernel<Kernel_traits, Is_dropout, Is_causal, Has_alibi, IsEvenNConst && IsEvenKConst, IsEvenKConst>;
// auto kernel = &flash_bwd_dq_dk_dv_loop_seqq_parallel_kernel<Kernel_traits, false, false, IsEvenNConst, IsEvenKConst>;
......
csrc/flash_attn/src/flash_fwd_kernel.h
View file @ 5ab9b366
......@@ -322,28 +322,14 @@ inline __device__ void compute_attn_1rowblock(const Params &params, const int bi
clear(acc_o);
float alibi_slope = !Has_alibi ? 0.0f : reinterpret_cast<float *>(params.alibi_slopes_ptr)[bidb * params.alibi_slopes_batch_stride + bidh] / params.scale_softmax;
// For performance reason, we separate out two kinds of iterations:
// those that need masking on S, and those that don't.
// We need masking on S for the very last block when K and V has length not multiple of kBlockN.
// We also need masking on S if it's causal, for the last ceil_div(kBlockM, kBlockN) blocks.
// We will have at least 1 "masking" iteration.
float alibi_slope = 0.0f;
if (Has_alibi) {
Tensor gAS = make_tensor(
make_gmem_ptr(
reinterpret_cast<ElementAccum *>(params.alibi_slopes_ptr)
+ bidb * params.alibi_slopes_batch_stride + bidh
),
Shape<_1>{});
Tensor rAS = make_fragment_like(gAS);
cute::copy(gAS, rAS);
alibi_slope = rAS(0);
// if (m_block == 0 && tidx == 0) {
// printf("%d,%d,%f\n", bidb, bidh, alibi_slope);
// }
}
// If not even_N, then seqlen_k might end in the middle of a block. In that case we need to
// mask 2 blocks (e.g. when kBlockM == kBlockN), not just 1.
constexpr int n_masking_steps = (!Is_causal && !Is_local)
......@@ -382,14 +368,13 @@ inline __device__ void compute_attn_1rowblock(const Params &params, const int bi
// can produce Inf / NaN.
if (Has_alibi) {
flash::apply_alibi(
flash::apply_alibi<Is_causal>(
scores,
n_block * kBlockN,
binfo.actual_seqlen_k,
m_block * kBlockM + (tidx / 32) * 16 + (tidx % 32) / 4,
binfo.actual_seqlen_q,
kNWarps * 16,
bidh, params.scale_softmax,
alibi_slope
);
}
......@@ -500,14 +485,13 @@ inline __device__ void compute_attn_1rowblock(const Params &params, const int bi
Tensor scores = make_tensor(acc_s.data(), flash::convert_layout_acc_rowcol(acc_s.layout()));
if (Has_alibi) {
flash::apply_alibi(
flash::apply_alibi<Is_causal>(
scores,
n_block * kBlockN,
binfo.actual_seqlen_k,
m_block * kBlockM + (tidx / 32) * 16 + (tidx % 32) / 4,
binfo.actual_seqlen_q,
kNWarps * 16,
bidh, params.scale_softmax,
alibi_slope
);
}
......@@ -950,28 +934,14 @@ inline __device__ void compute_attn_1rowblock_splitkv(const Params &params, cons
clear(acc_o);
float alibi_slope = !Has_alibi ? 0.0f : reinterpret_cast<float *>(params.alibi_slopes_ptr)[bidb * params.alibi_slopes_batch_stride + bidh] / params.scale_softmax;
// For performance reason, we separate out two kinds of iterations:
// those that need masking on S, and those that don't.
// We need masking on S for the very last block when K and V has length not multiple of kBlockN.
// We also need masking on S if it's causal, for the last ceil_div(kBlockM, kBlockN) blocks.
// We will have at least 1 "masking" iteration.
float alibi_slope = 0.0f;
if (Has_alibi) {
Tensor gAS = make_tensor(
make_gmem_ptr(
reinterpret_cast<ElementAccum *>(params.alibi_slopes_ptr)
+ bidb * params.alibi_slopes_batch_stride + bidh
),
Shape<_1>{});
Tensor rAS = make_fragment_like(gAS);
cute::copy(gAS, rAS);
alibi_slope = rAS(0);
// if (m_block == 0 && tidx == 0) {
// printf("%d,%d,%f\n", bidb, bidh, alibi_slope);
// }
}
// If not even_N, then seqlen_k might end in the middle of a block. In that case we need to
// mask 2 blocks (e.g. when kBlockM == kBlockN), not just 1.
constexpr int n_masking_steps = (!Is_causal && !Is_local)
......@@ -1006,14 +976,13 @@ inline __device__ void compute_attn_1rowblock_splitkv(const Params &params, cons
Tensor scores = make_tensor(acc_s.data(), flash::convert_layout_acc_rowcol(acc_s.layout()));
if (Has_alibi) {
flash::apply_alibi(
flash::apply_alibi<Is_causal>(
scores,
n_block * kBlockN,
binfo.actual_seqlen_k,
m_block * kBlockM + (tidx / 32) * 16 + (tidx % 32) / 4,
binfo.actual_seqlen_q,
kNWarps * 16,
bidh, params.scale_softmax,
alibi_slope
);
}
......@@ -1099,14 +1068,13 @@ inline __device__ void compute_attn_1rowblock_splitkv(const Params &params, cons
Tensor scores = make_tensor(acc_s.data(), flash::convert_layout_acc_rowcol(acc_s.layout()));
if (Has_alibi) {
flash::apply_alibi(
flash::apply_alibi<Is_causal>(
scores,
n_block * kBlockN,
binfo.actual_seqlen_k,
m_block * kBlockM + (tidx / 32) * 16 + (tidx % 32) / 4,
binfo.actual_seqlen_q,
kNWarps * 16,
bidh, params.scale_softmax,
alibi_slope
);
}
......
csrc/flash_attn/src/flash_fwd_launch_template.h
View file @ 5ab9b366
......@@ -45,7 +45,7 @@ void run_flash_fwd(Flash_fwd_params &params, cudaStream_t stream) {
BOOL_SWITCH(is_even_K, IsEvenKConst, [&] {
BOOL_SWITCH((params.window_size_left >= 0 || params.window_size_right >= 0) && !Is_causal, Is_local, [&] {
BOOL_SWITCH(return_softmax, ReturnSoftmaxConst, [&] {
BOOL_SWITCH(params.has_alibi, Has_alibi, [&] {
BOOL_SWITCH(params.alibi_slopes_ptr != nullptr, Has_alibi, [&] {
// Will only return softmax if dropout, to reduce compilation time.
// If not IsEvenKConst, we also set IsEvenMNConst to false to reduce number of templates.
// If return_softmax, set IsEvenMNConst to false to reduce number of templates
......@@ -86,7 +86,7 @@ void run_flash_splitkv_fwd(Flash_fwd_params &params, cudaStream_t stream) {
BOOL_SWITCH((params.window_size_left >= 0 || params.window_size_right >= 0) && !Is_causal, Is_local, [&] {
BOOL_SWITCH(params.num_splits > 1, Split, [&] {
BOOL_SWITCH(params.knew_ptr != nullptr, Append_KV, [&] {
BOOL_SWITCH(params.has_alibi, Has_alibi, [&] {
BOOL_SWITCH(params.alibi_slopes_ptr != nullptr, Has_alibi, [&] {
// If Append_KV, then we must have seqlen_offsets, which means cu_seqlens_k != nullptr.
// If not IsEvenKConst, we also set IsEvenMNConst to false to reduce number of templates.
// If Is_local, set Is_causal to false
......
csrc/flash_attn/src/softmax.h
View file @ 5ab9b366
......@@ -141,14 +141,12 @@ inline __device__ void apply_mask(Tensor<Engine, Layout> &tensor, const int max_
template <bool HasWSLeft=true, typename Engine, typename Layout>
inline __device__ void apply_mask_local(Tensor<Engine, Layout> &tensor, const int col_idx_offset_,
const int max_seqlen_k, const int row_idx_offset_,
const int max_seqlen_k, const int row_idx_offset,
const int max_seqlen_q, const int warp_row_stride,
const int window_size_left, const int window_size_right) {
// tensor has shape (ncol=(2, MMA_M), nrow=(2, MMA_N))
static_assert(Layout::rank == 2, "Only support 2D Tensor");
const int lane_id = threadIdx.x % 32;
// const int row_idx_offset = row_idx_offset_ + lane_id / 4;
const int row_idx_offset = row_idx_offset_;
const int col_idx_offset = col_idx_offset_ + (lane_id % 4) * 2;
#pragma unroll
for (int mi = 0; mi < size<0, 1>(tensor); ++mi) {
......@@ -180,17 +178,17 @@ inline __device__ void apply_mask_local(Tensor<Engine, Layout> &tensor, const in
template <typename Engine, typename Layout>
inline __device__ void apply_mask_causal(Tensor<Engine, Layout> &tensor, const int col_idx_offset_,
const int max_seqlen_k, const int row_idx_offset_,
const int max_seqlen_k, const int row_idx_offset,
const int max_seqlen_q, const int warp_row_stride) {
// Causal masking is equivalent to local masking with window_size_left = infinity and window_size_right = 0
apply_mask_local</*HasWSLeft=*/false>(tensor, col_idx_offset_, max_seqlen_k, row_idx_offset_,
apply_mask_local</*HasWSLeft=*/false>(tensor, col_idx_offset_, max_seqlen_k, row_idx_offset,
max_seqlen_q, warp_row_stride, -1, 0);
}
template <typename Engine0, typename Layout0, typename Engine1, typename Layout1>
inline __device__ void apply_mask_causal_w_idx(
Tensor<Engine0, Layout0> &tensor, Tensor<Engine1, Layout1> const &idx_rowcol,
const int col_idx_offset_, const int max_seqlen_k, const int row_idx_offset_)
const int col_idx_offset_, const int max_seqlen_k, const int row_idx_offset)
{
// tensor has shape (ncol=(2, MMA_M), nrow=(2, MMA_N))
static_assert(Layout0::rank == 2, "Only support 2D Tensor");
......@@ -199,7 +197,7 @@ inline __device__ void apply_mask_causal_w_idx(
CUTE_STATIC_ASSERT_V(size<1>(tensor) == size<1>(idx_rowcol));
#pragma unroll
for (int mi = 0; mi < size<0>(tensor); ++mi) {
const int col_idx_limit = std::min(max_seqlen_k, 1 + row_idx_offset_ + get<0>(idx_rowcol(mi, 0)));
const int col_idx_limit = std::min(max_seqlen_k, 1 + row_idx_offset + get<0>(idx_rowcol(mi, 0)));
#pragma unroll
for (int ni = 0; ni < size<1, 1>(tensor); ++ni) {
if (col_idx_offset_ + get<1>(idx_rowcol(0, ni)) >= col_idx_limit) {
......
flash_attn/flash_attn_interface.py
View file @ 5ab9b366
......@@ -53,12 +53,12 @@ def _flash_attn_forward(
k,
v,
None,
alibi_slopes,
dropout_p,
softmax_scale,
causal,
window_size[0],
window_size[1],
alibi_slopes,
return_softmax,
None,
)
......@@ -90,6 +90,7 @@ def _flash_attn_varlen_forward(
cu_seqlens_q,
cu_seqlens_k,
None,
alibi_slopes,
max_seqlen_q,
max_seqlen_k,
dropout_p,
......@@ -98,7 +99,6 @@ def _flash_attn_varlen_forward(
causal,
window_size[0],
window_size[1],
alibi_slopes,
return_softmax,
None,
)
......@@ -137,12 +137,12 @@ def _flash_attn_backward(
dq,
dk,
dv,
alibi_slopes,
dropout_p,
softmax_scale,
causal,
window_size[0],
window_size[1],
alibi_slopes,
None,
rng_state,
)
......@@ -185,6 +185,7 @@ def _flash_attn_varlen_backward(
dv,
cu_seqlens_q,
cu_seqlens_k,
alibi_slopes,
max_seqlen_q,
max_seqlen_k,
dropout_p,
......@@ -193,7 +194,6 @@ def _flash_attn_varlen_backward(
causal,
window_size[0],
window_size[1],
alibi_slopes,
None,
rng_state,
)
......@@ -613,6 +613,8 @@ def flash_attn_qkvpacked_func(
Default to 1 / sqrt(headdim).
causal: bool. Whether to apply causal attention mask (e.g., for auto-regressive modeling).
window_size: (left, right). If not (-1, -1), implements sliding window local attention.
alibi_slopes: (nheads,) or (batch_size, nheads), fp32. A bias of (-alibi_slope * |i - j|) is added to
the attention score of query i and key j.
return_attn_probs: bool. Whether to return the attention probabilities. This option is for
testing only. The returned probabilities are not guaranteed to be correct
(they might not have the right scaling).
......@@ -673,6 +675,9 @@ def flash_attn_kvpacked_func(
Default to 1 / sqrt(headdim).
causal: bool. Whether to apply causal attention mask (e.g., for auto-regressive modeling).
window_size: (left, right). If not (-1, -1), implements sliding window local attention.
alibi_slopes: (nheads,) or (batch_size, nheads), fp32. A bias of
(-alibi_slope * |i + seqlen_k - seqlen_q - j|)
is added to the attention score of query i and key j.
return_attn_probs: bool. Whether to return the attention probabilities. This option is for
testing only. The returned probabilities are not guaranteed to be correct
(they might not have the right scaling).
......@@ -732,6 +737,9 @@ def flash_attn_func(
Default to 1 / sqrt(headdim).
causal: bool. Whether to apply causal attention mask (e.g., for auto-regressive modeling).
window_size: (left, right). If not (-1, -1), implements sliding window local attention.
alibi_slopes: (nheads,) or (batch_size, nheads), fp32. A bias of
(-alibi_slope * |i + seqlen_k - seqlen_q - j|)
is added to the attention score of query i and key j.
return_attn_probs: bool. Whether to return the attention probabilities. This option is for
testing only. The returned probabilities are not guaranteed to be correct
(they might not have the right scaling).
......@@ -780,6 +788,8 @@ def flash_attn_varlen_qkvpacked_func(
Default to 1 / sqrt(headdim).
causal: bool. Whether to apply causal attention mask (e.g., for auto-regressive modeling).
window_size: (left, right). If not (-1, -1), implements sliding window local attention.
alibi_slopes: (nheads,) or (batch_size, nheads), fp32. A bias of (-alibi_slope * |i - j|)
is added to the attention score of query i and key j.
return_attn_probs: bool. Whether to return the attention probabilities. This option is for
testing only. The returned probabilities are not guaranteed to be correct
(they might not have the right scaling).
......@@ -858,6 +868,9 @@ def flash_attn_varlen_kvpacked_func(
Default to 1 / sqrt(headdim).
causal: bool. Whether to apply causal attention mask (e.g., for auto-regressive modeling).
window_size: (left, right). If not (-1, -1), implements sliding window local attention.
alibi_slopes: (nheads,) or (batch_size, nheads), fp32. A bias of
(-alibi_slope * |i + seqlen_k - seqlen_q - j|)
is added to the attention score of query i and key j.
return_attn_probs: bool. Whether to return the attention probabilities. This option is for
testing only. The returned probabilities are not guaranteed to be correct
(they might not have the right scaling).
......@@ -938,6 +951,9 @@ def flash_attn_varlen_func(
Default to 1 / sqrt(headdim).
causal: bool. Whether to apply causal attention mask (e.g., for auto-regressive modeling).
window_size: (left, right). If not (-1, -1), implements sliding window local attention.
alibi_slopes: (nheads,) or (batch_size, nheads), fp32. A bias of
(-alibi_slope * |i + seqlen_k - seqlen_q - j|)
is added to the attention score of query i and key j.
return_attn_probs: bool. Whether to return the attention probabilities. This option is for
testing only. The returned probabilities are not guaranteed to be correct
(they might not have the right scaling).
......@@ -981,8 +997,8 @@ def flash_attn_with_kvcache(
causal=False,
window_size=(-1, -1), # -1 means infinite context window
rotary_interleaved=True,
num_splits=0,
alibi_slopes=None,
num_splits=0,
):
"""
If k and v are not None, k_cache and v_cache will be updated *inplace* with the new values from
......@@ -1050,6 +1066,9 @@ def flash_attn_with_kvcache(
If True, rotary embedding will combine dimensions 0 & 1, 2 & 3, etc. If False,
rotary embedding will combine dimensions 0 & rotary_dim / 2, 1 & rotary_dim / 2 + 1
(i.e. GPT-NeoX style).
alibi_slopes: (nheads,) or (batch_size, nheads), fp32. A bias of
(-alibi_slope * |i + seqlen_k - seqlen_q - j|)
is added to the attention score of query i and key j.
num_splits: int. If > 1, split the key/value into this many chunks along the sequence.
If num_splits == 1, we don't split the key/value. If num_splits == 0, we use a heuristic
to automatically determine the number of splits.
......@@ -1080,6 +1099,7 @@ def flash_attn_with_kvcache(
rotary_cos,
rotary_sin,
cache_batch_idx,
alibi_slopes,
None,
softmax_scale,
causal,
......@@ -1087,6 +1107,5 @@ def flash_attn_with_kvcache(
window_size[1],
rotary_interleaved,
num_splits,
alibi_slopes,
)
return out
tests/test_alibi.py deleted 100644 → 0
View file @ bc28eacc
import math
import pytest
import torch
import torch.nn.functional as F
from einops import rearrange, repeat
from flash_attn import (flash_attn_func, flash_attn_kvpacked_func,
flash_attn_qkvpacked_func, flash_attn_varlen_func,
flash_attn_varlen_kvpacked_func,
flash_attn_varlen_qkvpacked_func,
flash_attn_with_kvcache)
from flash_attn.bert_padding import pad_input, unpad_input
from flash_attn.flash_attn_interface import _get_block_size
from flash_attn.flash_attn_triton import \
flash_attn_func as flash_attn_func_triton
from flash_attn.layers.rotary import apply_rotary_emb
MAX_HEADDIM_SM8x = 192
is_sm75 = torch.cuda.get_device_capability("cuda") == (7, 5)
is_sm8x = torch.cuda.get_device_capability("cuda")[0] == 8
is_sm80 = torch.cuda.get_device_capability("cuda") == (8, 0)
is_sm90 = torch.cuda.get_device_capability("cuda") == (9, 0)
def generate_alibi(max_seq_len, num_attention_heads, tp_world_size, tp_index, key_padding_mask=None, device="cuda"):
def get_slopes(n):
def get_slopes_power_of_2(n):
start = (2 ** (-2 ** -(math.log2(n) - 3)))
ratio = start
return [start * ratio ** i for i in range(n)]
if math.log2(n).is_integer():
return get_slopes_power_of_2(n)
else:
closest_power_of_2 = 2 ** math.floor(math.log2(n))
return get_slopes_power_of_2(closest_power_of_2) + get_slopes(2 * closest_power_of_2)[0::2][
:n - closest_power_of_2]
slopes = torch.tensor(get_slopes(num_attention_heads)).to(device=device)
# Select the part of the tensor that corresponds to our tensor parallel index.
assert (num_attention_heads/tp_world_size).is_integer(
), "it works only when (num_attention_heads/tp_world_size) is integer"
nh_tp = num_attention_heads // tp_world_size
slopes = slopes[nh_tp * tp_index:nh_tp * (tp_index + 1)]
if (key_padding_mask is None):
arange_tensor = rearrange(torch.arange(max_seq_len), "sqk -> 1 sqk").to(device=device)
else:
arange_tensor = (key_padding_mask.cumsum(dim=-1, dtype=slopes.dtype) - 1) \
.masked_fill_(~key_padding_mask, torch.finfo(torch.float).min).to(device=device)
arange_tensor = rearrange(arange_tensor, 'b sqk -> b 1 1 sqk')
# (1, nheads, 1, seqlen_k) or (batch, nheads, 1, seqlen_k)
alibi_tensor = rearrange(slopes, 'nh -> 1 nh 1 1') * arange_tensor
return alibi_tensor, slopes
def generate_random_padding_mask(max_seqlen, batch_size, device, mode="random", right_padding=True):
assert mode in ["full", "random", "third"]
if mode == "full":
lengths = torch.full((batch_size, 1), max_seqlen,
device=device, dtype=torch.int32)
elif mode == "random":
lengths = torch.randint(
max(1, max_seqlen - 20), max_seqlen + 1, (batch_size, 1), device=device
)
elif mode == "third":
lengths = torch.randint(
max_seqlen // 3, max_seqlen + 1, (batch_size, 1), device=device)
if right_padding:
padding_mask = (
repeat(torch.arange(max_seqlen, device=device),
"s -> b s", b=batch_size) < lengths
)
else:
padding_mask = (
repeat(torch.arange(start=max_seqlen-1, end=-1, step=-1, device=device),
"s -> b s", b=batch_size) < lengths
)
return padding_mask
def generate_qkv(
q, k, v, query_padding_mask=None, key_padding_mask=None, kvpacked=False, qkvpacked=False
):
"""
Arguments:
q: (batch_size, seqlen_q, nheads, d)
k: (batch_size, seqlen_k, nheads_k, d)
v: (batch_size, seqlen_k, nheads_k, d)
query_padding_mask: (batch_size, seqlen), bool
key_padding_mask: (batch_size, seqlen), bool
"""
assert not (kvpacked and qkvpacked)
batch_size, seqlen_q, nheads, d = q.shape
_, seqlen_k, nheads_k, _ = k.shape
assert k.shape == (batch_size, seqlen_k, nheads_k, d)
assert v.shape == (batch_size, seqlen_k, nheads_k, d)
if query_padding_mask is not None:
q_unpad, indices_q, cu_seqlens_q, max_seqlen_q = unpad_input(
q, query_padding_mask)
def output_pad_fn(output_unpad): return pad_input(
output_unpad, indices_q, batch_size, seqlen_q
)
else:
q_unpad = rearrange(q, "b s h d -> (b s) h d")
cu_seqlens_q = torch.arange(
0, (batch_size + 1) * seqlen_q, step=seqlen_q, dtype=torch.int32, device=q_unpad.device
)
max_seqlen_q = seqlen_q
def output_pad_fn(output_unpad): return rearrange(
output_unpad, "(b s) h d -> b s h d", b=batch_size
)
if key_padding_mask is not None:
k_unpad, indices_k, cu_seqlens_k, max_seqlen_k = unpad_input(
k, key_padding_mask)
v_unpad, _, _, _ = unpad_input(v, key_padding_mask)
else:
k_unpad = rearrange(k, "b s h d -> (b s) h d")
v_unpad = rearrange(v, "b s h d -> (b s) h d")
cu_seqlens_k = torch.arange(
0, (batch_size + 1) * seqlen_k, step=seqlen_k, dtype=torch.int32, device=k_unpad.device
)
max_seqlen_k = seqlen_k
if qkvpacked:
assert (query_padding_mask == key_padding_mask).all()
assert nheads == nheads_k
qkv_unpad = torch.stack([q_unpad, k_unpad, v_unpad], dim=1)
qkv = torch.stack([q, k, v], dim=2)
if query_padding_mask is not None:
def dqkv_pad_fn(dqkv_unpad): return pad_input(
dqkv_unpad, indices_q, batch_size, seqlen_q)
else:
def dqkv_pad_fn(dqkv_unpad): return rearrange(
dqkv_unpad, "(b s) t h d -> b s t h d", b=batch_size
)
return (
qkv_unpad.detach().requires_grad_(),
cu_seqlens_q,
max_seqlen_q,
qkv.detach().requires_grad_(),
output_pad_fn,
dqkv_pad_fn,
)
elif kvpacked:
kv_unpad = torch.stack([k_unpad, v_unpad], dim=1)
kv = torch.stack([k, v], dim=2)
dq_pad_fn = output_pad_fn
if key_padding_mask is not None:
def dkv_pad_fn(dkv_unpad): return pad_input(
dkv_unpad, indices_k, batch_size, seqlen_k)
else:
def dkv_pad_fn(dkv_unpad): return rearrange(
dkv_unpad, "(b s) t h d -> b s t h d", b=batch_size
)
return (
q_unpad.detach().requires_grad_(),
kv_unpad.detach().requires_grad_(),
cu_seqlens_q,
cu_seqlens_k,
max_seqlen_q,
max_seqlen_k,
q.detach().requires_grad_(),
kv.detach().requires_grad_(),
output_pad_fn,
dq_pad_fn,
dkv_pad_fn,
)
else:
dq_pad_fn = output_pad_fn
if key_padding_mask is not None:
def dk_pad_fn(dk_unpad): return pad_input(
dk_unpad, indices_k, batch_size, seqlen_k)
else:
def dk_pad_fn(dk_unpad): return rearrange(
dk_unpad, "(b s) h d -> b s h d", b=batch_size)
return (
q_unpad.detach().requires_grad_(),
k_unpad.detach().requires_grad_(),
v_unpad.detach().requires_grad_(),
cu_seqlens_q,
cu_seqlens_k,
max_seqlen_q,
max_seqlen_k,
q.detach().requires_grad_(),
k.detach().requires_grad_(),
v.detach().requires_grad_(),
output_pad_fn,
dq_pad_fn,
dk_pad_fn,
)
def construct_local_mask(
seqlen_q,
seqlen_k,
window_size=(-1, -1), # -1 means infinite window size
query_padding_mask=None,
key_padding_mask=None,
device=None,
):
row_idx = rearrange(torch.arange(
seqlen_q, device=device, dtype=torch.long), "s -> s 1")
col_idx = torch.arange(seqlen_k, device=device, dtype=torch.long)
sk = (
seqlen_k
if key_padding_mask is None
else rearrange(key_padding_mask.sum(-1), "b -> b 1 1 1")
)
sq = (
seqlen_q
if query_padding_mask is None
else rearrange(query_padding_mask.sum(-1), "b -> b 1 1 1")
)
if window_size[0] < 0:
return col_idx > row_idx + sk - sq + window_size[1]
else:
sk = torch.full_like(
col_idx, seqlen_k) if key_padding_mask is None else sk
return torch.logical_or(
col_idx > torch.minimum(row_idx + sk - sq + window_size[1], sk),
col_idx < row_idx + sk - sq - window_size[0],
)
def attention_ref(
q,
k,
v,
query_padding_mask=None,
key_padding_mask=None,
dropout_p=0.0,
dropout_mask=None,
causal=False,
window_size=(-1, -1), # -1 means infinite window size
upcast=True,
reorder_ops=False,
bias=None
):
"""
Arguments:
q: (batch_size, seqlen_q, nheads, head_dim)
k: (batch_size, seqlen_k, nheads_k, head_dim)
v: (batch_size, seqlen_k, nheads_k, head_dim)
query_padding_mask: (batch_size, seqlen_q)
key_padding_mask: (batch_size, seqlen_k)
dropout_p: float
dropout_mask: (batch_size, nheads, seqlen_q, seqlen_k)
causal: whether to apply causal masking
window_size: (int, int), left and right window size
upcast: whether to cast all inputs to fp32, do all computation in fp32, then cast
output back to fp16/bf16.
reorder_ops: whether to change the order of operations (scaling k instead of scaling k, etc.)
without changing the math. This is to estimate the numerical error from operation
reordering.
Output:
output: (batch_size, seqlen_q, nheads, head_dim)
attention: (batch_size, nheads, seqlen_q, seqlen_k), softmax after dropout
"""
if causal:
window_size = (window_size[0], 0)
dtype_og = q.dtype
if upcast:
q, k, v = q.float(), k.float(), v.float()
seqlen_q, seqlen_k = q.shape[1], k.shape[1]
k = repeat(k, "b s h d -> b s (h g) d", g=q.shape[2] // k.shape[2])
v = repeat(v, "b s h d -> b s (h g) d", g=q.shape[2] // v.shape[2])
d = q.shape[-1]
if not reorder_ops:
scores = torch.einsum("bthd,bshd->bhts", q / math.sqrt(d), k)
else:
scores = torch.einsum("bthd,bshd->bhts", q, k / math.sqrt(d))
if bias is not None:
bias = bias.to(scores.dtype)
scores += bias
if key_padding_mask is not None:
scores.masked_fill_(rearrange(~key_padding_mask,
"b s -> b 1 1 s"), float("-inf"))
if window_size[0] >= 0 or window_size[1] >= 0:
local_mask = construct_local_mask(
seqlen_q,
seqlen_k,
window_size,
query_padding_mask,
key_padding_mask,
q.device,
)
scores.masked_fill_(local_mask, float("-inf"))
attention = torch.softmax(scores, dim=-1)
# Some rows might be completely masked out so we fill them with zero instead of NaN
if window_size[0] >= 0 or window_size[1] >= 0:
attention = attention.masked_fill(
torch.all(local_mask, dim=-1, keepdim=True), 0.0)
# We want to mask here so that the attention matrix doesn't have any NaNs
# Otherwise we'll get NaN in dV
if query_padding_mask is not None:
attention = attention.masked_fill(
rearrange(~query_padding_mask, "b s -> b 1 s 1"), 0.0)
dropout_scaling = 1.0 / (1 - dropout_p)
# attention_drop = attention.masked_fill(~dropout_mask, 0.0) * dropout_scaling
# output = torch.einsum('bhts,bshd->bthd', attention_drop , v)
if dropout_mask is not None:
attention_drop = attention.masked_fill(~dropout_mask, 0.0)
else:
attention_drop = attention
output = torch.einsum(
"bhts,bshd->bthd", attention_drop, v * dropout_scaling)
if query_padding_mask is not None:
output.masked_fill_(
rearrange(~query_padding_mask, "b s -> b s 1 1"), 0.0)
return output.to(dtype=dtype_og), attention.to(dtype=dtype_og)
def attention_kvpacked_ref(
q,
kv,
query_padding_mask=None,
key_padding_mask=None,
dropout_p=0.0,
dropout_mask=None,
causal=False,
window_size=(-1, -1), # -1 means infinite window size
upcast=True,
reorder_ops=False,
):
return attention_ref(
q,
kv[:, :, 0],
kv[:, :, 1],
query_padding_mask,
key_padding_mask,
dropout_p,
dropout_mask,
upcast=upcast,
causal=causal,
window_size=window_size,
reorder_ops=reorder_ops,
)
def attention_qkvpacked_ref(
qkv,
key_padding_mask=None,
dropout_p=0.0,
dropout_mask=None,
causal=False,
window_size=(-1, -1), # -1 means infinite window size
upcast=True,
reorder_ops=False,
):
return attention_ref(
qkv[:, :, 0],
qkv[:, :, 1],
qkv[:, :, 2],
key_padding_mask,
key_padding_mask,
dropout_p,
dropout_mask,
upcast=upcast,
causal=causal,
window_size=window_size,
reorder_ops=reorder_ops,
)
def generate_sparsity_mask(seqlen, sparsity=0.3):
repeats = seqlen // 16 // 2
# mask = torch.stack([torch.tensor([1, 0] * repeats, dtype=torch.bool, device='cuda'),
# torch.tensor([0, 1] * repeats, dtype=torch.bool, device='cuda')], dim=-1)
# mask = torch.stack([torch.tensor([1, 1] * repeats, dtype=torch.bool, device='cuda'),
# torch.tensor([1, 1] * repeats, dtype=torch.bool, device='cuda')], dim=-1)
# mask = torch.stack([torch.tensor([1, 1] * repeats, dtype=torch.bool, device='cuda')], dim=-1)
# mask = torch.stack([torch.tensor([1, 0] * repeats, dtype=torch.bool, device='cuda')], dim=-1)
nrow, ncol = seqlen // 16, seqlen // 256
mask = torch.rand(nrow, ncol, device="cuda") < sparsity
return mask
def attention_blocksparse_ref(qkv, blockmask, attn_mask, dropout_p, dropout_mask):
"""
Arguments:
qkv: (batch_size, seqlen, 3, nheads, head_dim)
blockmask: (seqlen / 16, seqlen / 256)
attn_mask: (batch_size, seqlen)
dropout_p: float
dropout_mask: (batch_size, nheads, seqlen, seqlen)
Output:
output: (batch_size, seqlen, nheads, head_dim)
attention: softmax after dropout
"""
q, k, v = qkv.float().unbind(dim=2)
d = qkv.shape[-1]
seqlen = qkv.shape[1]
scores = torch.einsum("bthd,bshd->bhts", q / math.sqrt(d), k)
scores.masked_fill_(rearrange(~attn_mask, "b s -> b 1 1 s"), float("-inf"))
blockmask = repeat(blockmask, "s_16 s_256 -> (s_16 16) (s_256 256)")
blockmask = blockmask[:seqlen, :seqlen]
scores.masked_fill_(rearrange(~blockmask, "t s -> 1 1 t s"), float("-inf"))
attention = torch.softmax(scores, dim=-1)
attention = attention.masked_fill(
rearrange(~attn_mask, "b s -> b 1 s 1"), 0.0)
attention = attention.masked_fill_(
rearrange(~blockmask, "t s -> 1 1 t s"), 0.0)
attention_drop = attention.masked_fill(
~dropout_mask, 0.0) / (1 - dropout_p)
output = torch.einsum("bhts,bshd->bthd", attention_drop, v)
output.masked_fill_(rearrange(~attn_mask, "b s -> b s 1 1"), 0)
return output.to(dtype=qkv.dtype), attention.to(dtype=qkv.dtype)
def convert_flash_attn_S_to_softmax(
S,
seqlen_q,
seqlen_k,
query_padding_mask,
key_padding_mask,
head_dim,
is_dropout,
causal=False,
window_size=(-1, -1), # -1 means infinite window size
):
"""FlashAttention stores the S matrix in a different way.
Arguments:
S: (batch_size, nheads, seqlen_q_rounded, seqlen_k_rounded)
query_padding_mask: (batch_size, seqlen_q_rounded)
key_padding_mask: (batch_size, seqlen_k_rounded)
"""
if causal:
window_size = (window_size[0], 0)
seqlen_q_rounded, seqlen_k_rounded = S.shape[-2:]
warps_n = 4
blocksize_m, blocksize_n = _get_block_size(
S.device, head_dim, is_dropout, causal)
nblocks_n = (seqlen_k_rounded + blocksize_n - 1) // blocksize_n
nblocks_m = (seqlen_q_rounded + blocksize_m - 1) // blocksize_m
mmas_n = (blocksize_n + 16 - 1) // 16
S_flat = rearrange(
S,
"b h (nblocks_m blocksize_m) (nblocks_n blocksize_n) -> b h nblocks_m nblocks_n (blocksize_m blocksize_n)",
blocksize_m=blocksize_m,
blocksize_n=blocksize_n,
)
S_converted = rearrange(
S_flat,
"b h nblocks_m nblocks_n (mmas_n mmas_m warps_n eight four c2 c1 c0) -> b h (nblocks_m mmas_m warps_n c1 eight) (nblocks_n mmas_n c2 four c0)",
mmas_n=mmas_n,
warps_n=warps_n,
eight=8,
c0=2,
c1=2,
c2=2,
four=4,
)
if window_size[0] >= 0 or window_size[1] >= 0:
local_mask = construct_local_mask(
seqlen_q,
seqlen_k,
window_size,
query_padding_mask,
key_padding_mask,
S.device,
)
local_mask = F.pad(
local_mask,
(0, seqlen_k_rounded - seqlen_k, 0, seqlen_q_rounded - seqlen_q),
value=True,
)
S_converted.masked_fill_(local_mask, 0.0)
# Need to zero out things not in attention_mask in case S was initialized with random values
# and some of those values aren't overwritten.
seqlen_q_og = (
query_padding_mask.shape[-1] if query_padding_mask is not None else seqlen_q_rounded
)
if query_padding_mask is not None:
query_padding_mask = F.pad(
query_padding_mask, (0, seqlen_q_rounded - seqlen_q_og))
S_converted = S_converted.masked_fill(
rearrange(~query_padding_mask, "b s -> b 1 s 1"), 0.0)
seqlen_k_og = key_padding_mask.shape[-1] if key_padding_mask is not None else seqlen_k
if key_padding_mask is not None:
key_padding_mask = F.pad(
key_padding_mask, (0, seqlen_k_rounded - seqlen_k_og))
S_converted = S_converted.masked_fill(
rearrange(~key_padding_mask, "b s -> b 1 1 s"), 0.0)
S_converted = F.pad(S_converted, (0, 0, 0, seqlen_q_og - seqlen_q_rounded))
S_converted = F.pad(S_converted, (0, seqlen_k_og - seqlen_k_rounded))
return S_converted[:, :, :seqlen_q, :seqlen_k]
def normalize_flash_attn_S(
attn_unnorm,
q,
k,
v,
query_padding_mask=None,
key_padding_mask=None,
is_dropout=False,
causal=False,
window_size=(-1, -1), # -1 means infinite window size
):
"""
Arguments:
q: (batch_size, seqlen_q, nheads, head_dim)
k, v: (batch_size, seqlen_k, nheads, head_dim)
key_padding_mask: (batch_size, seqlen_q)
Output:
softmax_lse: (batch_size, nheads, seqlen_q)
softmax_max: (batch_size, nheads, seqlen_q)
"""
if causal:
window_size = (window_size[0], 0)
q, k, v = q.float(), k.float(), v.float()
_, seqlen_q, _, head_dim = q.shape
seqlen_k = k.shape[1]
scores = torch.einsum("bthd,bshd->bhts", q / math.sqrt(head_dim), k)
if key_padding_mask is not None:
scores.masked_fill_(rearrange(~key_padding_mask,
"b s -> b 1 1 s"), float("-inf"))
if window_size[0] >= 0 or window_size[1] >= 0:
local_mask = construct_local_mask(
seqlen_q,
seqlen_k,
window_size,
query_padding_mask,
key_padding_mask,
q.device,
)
scores.masked_fill_(local_mask, float("-inf"))
_, block_size_n = _get_block_size(
scores.device, head_dim, is_dropout, causal)
scores_block = scores.split(block_size_n, dim=-1)
lse_block = torch.stack([torch.logsumexp(s, dim=-1)
for s in scores_block], dim=-1)
lse = torch.logsumexp(lse_block, dim=-1)
# lse could be -inf (i.e. all values in scores are -inf), and we want to set those to inf
# so that when we do torch.exp(m - lse), we get 0.0 instead of NaN.
lse[lse == float("-inf")] = float("inf")
scores_max_block = torch.stack(
[torch.amax(s, dim=-1) for s in scores_block], dim=-1)
cummax_block = torch.cummax(
scores_max_block.flip(-1), dim=-1).values.flip(-1).unbind(dim=-1)
attn_unnorm_block = attn_unnorm.split(block_size_n, dim=-1)
attn_norm = torch.cat(
[
a * rearrange(torch.exp(m - lse), "b h s -> b h s 1")
for a, m in zip(attn_unnorm_block, cummax_block)
],
dim=-1,
)
if query_padding_mask is not None:
attn_norm.masked_fill_(
rearrange(~query_padding_mask, "b s -> b 1 s 1"), 0.0)
return attn_norm.to(dtype=attn_unnorm.dtype)
def get_dropout_fraction(
dropout_mask,
query_padding_mask=None,
key_padding_mask=None,
causal=False,
window_size=(-1, -1), # -1 means infinite window size
):
"""
dropout_mask: (batch_size, nheads, seqlen_q, seqlen_k), bool. True means keep, False means drop.
query_padding_mask: (batch_size, seqlen_q)
key_padding_mask: (batch_size, seqlen_k)
"""
if causal:
window_size = (window_size[0], 0)
batch_size, nheads, seqlen_q, seqlen_k = dropout_mask.shape
dropped = ~dropout_mask
valid = torch.ones_like(dropout_mask)
if query_padding_mask is not None:
dropped.masked_fill_(
rearrange(~query_padding_mask, "b s -> b 1 s 1"), False)
valid.masked_fill_(
rearrange(~query_padding_mask, "b s -> b 1 s 1"), False)
if key_padding_mask is not None:
dropped.masked_fill_(
rearrange(~key_padding_mask, "b s -> b 1 1 s"), False)
valid.masked_fill_(
rearrange(~key_padding_mask, "b s -> b 1 1 s"), False)
if window_size[0] >= 0 or window_size[1] >= 0:
local_mask = construct_local_mask(
seqlen_q,
seqlen_k,
window_size,
query_padding_mask,
key_padding_mask,
dropout_mask.device,
)
dropped.masked_fill_(local_mask, False)
valid.masked_fill_(local_mask, False)
dropped_total = dropped.sum()
return dropped.sum() / valid.sum()
@pytest.mark.parametrize(
"dtype", [torch.float16]
)
@pytest.mark.parametrize(
"b_sq",
[
(32, 512),
(16, 1024),
(8, 2048),
(4, 4096),
(2, 8192),
(1, 16384)
]
)
@pytest.mark.parametrize(
"nh_hd",
[
(32, 64),
(16, 128),
(40, 128) # non power of 2 nh
]
)
@pytest.mark.parametrize(
"tp_world_size", [1, 2, 4]
)
def test_flash_attn_func(b_sq, nh_hd, tp_world_size, dtype):
b, sq = b_sq
nh, hd = nh_hd
nh_tp = nh // tp_world_size
q, k, v = [torch.randn(b, sq, nh_tp, hd, device="cuda",
dtype=dtype, requires_grad=True) for _ in range(3)]
dout = torch.rand_like(q)
for tp_index in range(tp_world_size):
alibi, alibi_slopes = generate_alibi(
max_seq_len=sq,
num_attention_heads=nh,
tp_world_size=tp_world_size,
tp_index=tp_index,
key_padding_mask=None,
device="cuda"
)
triton_out = flash_attn_func_triton(
q, k, v, alibi, True, hd**(-0.5))
triton_out.backward(dout)
triton_dq, q.grad = q.grad.clone(), None
triton_dk, k.grad = k.grad.clone(), None
triton_dv, v.grad = v.grad.clone(), None
flash_out = flash_attn_func(q, k, v, causal=True, alibi_slopes=repeat(alibi_slopes, "nh -> b nh", b=b))
flash_out.backward(dout)
flash_dq, q.grad = q.grad.clone(), None
flash_dk, k.grad = k.grad.clone(), None
flash_dv, v.grad = v.grad.clone(), None
assert torch.allclose(flash_out, triton_out, atol=1e-2, rtol=0.)
assert torch.allclose(flash_dq, triton_dq, atol=1e-2, rtol=0.)
assert torch.allclose(flash_dk, triton_dk, atol=1e-2, rtol=0.)
assert torch.allclose(flash_dv, triton_dv, atol=1e-2, rtol=0.)
@pytest.mark.parametrize(
"dtype", [torch.float16]
)
@pytest.mark.parametrize(
"right_padding", [True, False]
)
@pytest.mark.parametrize(
"b_sq",
[
(32, 512),
(16, 1024),
(8, 2048),
(4, 4096),
(2, 8192),
(1, 16384)
]
)
@pytest.mark.parametrize(
"nh_hd",
[
(32, 64),
(16, 128),
(40, 128) # non power of 2 nh
]
)
@pytest.mark.parametrize(
"tp_world_size", [1, 2, 4]
)
def test_flash_attn_varlen_func(b_sq, nh_hd, tp_world_size, right_padding, dtype):
b, sqk = b_sq
nh, hd = nh_hd
nh_tp = nh // tp_world_size
# flash_attn_func_triton(), flash-attention v2 (above v2.1) causal logic are different
# so only (seqlen_q == 1, causal=False to triton ver.) shows correct results
# https://github.com/huggingface/text-generation-inference/blob/v1.1.1/server/text_generation_server/models/custom_modeling/mpt_modeling.py#L53-L63
q = torch.randn(b, 1, nh_tp, hd, device="cuda", dtype=dtype, requires_grad=True)
k, v = [torch.randn(b, sqk, nh_tp, hd, device="cuda",
dtype=dtype, requires_grad=True) for _ in range(2)]
dout = torch.rand_like(q)
padding_mask = generate_random_padding_mask(sqk, b, "cuda", "random", right_padding)
(
q_unpad,
k_unpad,
v_unpad,
cu_seqlens_q,
cu_seqlens_k,
max_seqlen_q,
max_seqlen_k,
q,
k,
v,
output_pad_fn,
dq_pad_fn,
dk_pad_fn,
) = generate_qkv(q, k, v, None, padding_mask, kvpacked=False)
for tp_index in range(tp_world_size):
alibi, alibi_slopes = generate_alibi(
max_seq_len=sqk,
num_attention_heads=nh,
tp_world_size=tp_world_size,
tp_index=tp_index,
key_padding_mask=padding_mask,
device="cuda"
)
triton_out = flash_attn_func_triton(
q, k, v, alibi, False, hd**(-0.5))
triton_out.backward(dout)
triton_dq, q.grad = q.grad.clone(), None
triton_dk, k.grad = k.grad.clone(), None
triton_dv, v.grad = v.grad.clone(), None
flash_out_unpad = flash_attn_varlen_func(
q_unpad,
k_unpad,
v_unpad,
cu_seqlens_q,
cu_seqlens_k,
max_seqlen_q,
max_seqlen_k,
causal=True,
alibi_slopes=repeat(alibi_slopes, "nh -> b nh", b=b)
)
flash_out = output_pad_fn(flash_out_unpad)
flash_out.backward(dout)
flash_dq_unpad, q_unpad.grad = q_unpad.grad.clone(), None
flash_dk_unpad, k_unpad.grad = k_unpad.grad.clone(), None
flash_dv_unpad, v_unpad.grad = v_unpad.grad.clone(), None
flash_dq = dq_pad_fn(flash_dq_unpad)
flash_dk = dk_pad_fn(flash_dk_unpad)
flash_dv = dk_pad_fn(flash_dv_unpad)
assert torch.allclose(flash_out, triton_out, atol=1e-2, rtol=0.)
assert torch.allclose(flash_dq, triton_dq, atol=1e-2, rtol=0.)
assert torch.allclose(flash_dk, triton_dk, atol=1e-2, rtol=0.)
assert torch.allclose(flash_dv, triton_dv, atol=1e-2, rtol=0.)
@pytest.mark.parametrize("alibi", [True])
@pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
# @pytest.mark.parametrize("dtype", [torch.float16])
@pytest.mark.parametrize("num_splits", [1, 0])
# @pytest.mark.parametrize("num_splits", [0])
@pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"])
# @pytest.mark.parametrize("mha_type", ["mha"])
@pytest.mark.parametrize("new_kv", [False, True])
# @pytest.mark.parametrize("new_kv", [True])
# @pytest.mark.parametrize("local", [False, True])
@pytest.mark.parametrize("local", [False])
# @pytest.mark.parametrize("causal", [False, True])
@pytest.mark.parametrize("causal", [True])
@pytest.mark.parametrize("seqlen_new_eq_seqlen_q", [True, False])
# @pytest.mark.parametrize("seqlen_new_eq_seqlen_q", [True])
@pytest.mark.parametrize("rotary_interleaved", [False, True])
# @pytest.mark.parametrize("rotary_interleaved", [False])
@pytest.mark.parametrize("rotary_fraction", [0.0, 0.5, 1.0])
# @pytest.mark.parametrize("rotary_fraction", [0.0])
@pytest.mark.parametrize("has_batch_idx", [False, True])
# @pytest.mark.parametrize("has_batch_idx", [True])
@pytest.mark.parametrize("d", [32, 59, 64, 80, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize('d', [32, 64, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize('d', [32, 40, 64, 80, 96, 128, 160, 192])
# @pytest.mark.parametrize('d', [56, 80])
# @pytest.mark.parametrize("d", [128])
@pytest.mark.parametrize(
"seqlen_q,seqlen_k",
[
(1, 128),
(1, 339),
(3, 1024),
(64, 800),
(64, 256),
(3, 799),
(64, 2048),
(16, 20000),
(1, 128 * 1024),
(16, 128 * 1024),
(128, 128),
],
)
# @pytest.mark.parametrize('seqlen_q,seqlen_k', [(256, 128)])
def test_flash_attn_kvcache(
seqlen_q,
seqlen_k,
d,
has_batch_idx,
rotary_fraction,
rotary_interleaved,
seqlen_new_eq_seqlen_q,
causal,
local,
new_kv,
mha_type,
num_splits,
dtype,
alibi,
):
if seqlen_q > seqlen_k and new_kv:
pytest.skip()
if not new_kv and rotary_fraction > 0.0:
pytest.skip()
device = "cuda"
# set seed
torch.random.manual_seed(0)
batch_size = 2
batch_size_cache = batch_size if not has_batch_idx else batch_size * 2
nheads = 8
# rotary_dim must be a multiple of 16, and must be <= d
rotary_dim = math.floor(int(rotary_fraction * d) / 16) * 16
nheads_k = nheads if mha_type == "mha" else (1 if mha_type == "mqa" else 4)
assert nheads % nheads_k == 0
window_size = (-1, -1) if not local else torch.randint(0, seqlen_k, (2,))
q = torch.randn(batch_size, seqlen_q, nheads,
d, device=device, dtype=dtype)
seqlen_new = seqlen_q if seqlen_new_eq_seqlen_q else torch.randint(
1, seqlen_q + 1, (1,)).item()
if new_kv:
k = torch.randn(batch_size, seqlen_new, nheads_k,
d, device=device, dtype=dtype)
v = torch.randn(batch_size, seqlen_new, nheads_k,
d, device=device, dtype=dtype)
else:
k, v = None, None
k_cache = torch.randn(batch_size_cache, seqlen_k,
nheads_k, d, device=device, dtype=dtype)
v_cache = torch.randn(batch_size_cache, seqlen_k,
nheads_k, d, device=device, dtype=dtype)
cache_seqlens = torch.randint(
0,
# If we don't use seqlen_q in the case of causal and rotary, cos/sin won't be long enough
(seqlen_k - (seqlen_q if (causal or local)
and rotary_dim > 1 else seqlen_new) + 1)
if new_kv
else (seqlen_k + 1),
(batch_size,),
dtype=torch.int32,
device=device,
)
if has_batch_idx:
cache_batch_idx = torch.randperm(
batch_size_cache, dtype=torch.int32, device=device)[:batch_size]
else:
cache_batch_idx = None
# cache_seqlens = torch.tensor([64], dtype=torch.int32, device=device)
if rotary_dim > 0:
angle = torch.rand(seqlen_k, rotary_dim // 2,
device=device) * 2 * math.pi
cos = torch.cos(angle).to(dtype=dtype)
sin = torch.sin(angle).to(dtype=dtype)
if causal or local:
q_ro = apply_rotary_emb(
q, cos, sin, seqlen_offsets=cache_seqlens, interleaved=rotary_interleaved
)
else:
q_ro = rearrange(
apply_rotary_emb(
rearrange(q, "b s h d -> b 1 (s h) d"),
cos,
sin,
seqlen_offsets=cache_seqlens,
interleaved=rotary_interleaved,
),
"b 1 (s h) d -> b s h d",
s=seqlen_q,
)
# q_ro = q
k_ro = apply_rotary_emb(
k, cos, sin, seqlen_offsets=cache_seqlens, interleaved=rotary_interleaved
)
else:
cos, sin = None, None
q_ro, k_ro = q, k
# k_cache[:, 64:] = -1
k_cache_ref = (
k_cache if not has_batch_idx else k_cache[cache_batch_idx]).clone()
v_cache_ref = (
v_cache if not has_batch_idx else v_cache[cache_batch_idx]).clone()
arange = rearrange(torch.arange(seqlen_k, device=device), "s -> 1 s")
cache_seqlens_expanded = rearrange(cache_seqlens, "b -> b 1")
if new_kv:
update_mask = torch.logical_and(
cache_seqlens_expanded <= arange, arange < cache_seqlens_expanded + seqlen_new
)
k_cache_ref[update_mask] = rearrange(k_ro, "b s ... -> (b s) ...")
v_cache_ref[update_mask] = rearrange(v, "b s ... -> (b s) ...")
k_cache_rep = repeat(
k_cache_ref, "b s h d -> b s (h g) d", g=nheads // nheads_k)
v_cache_rep = repeat(
v_cache_ref, "b s h d -> b s (h g) d", g=nheads // nheads_k)
if alibi:
seqlen_alibi = k_cache_rep.shape[1]
alibi_tensor, alibi_slopes = generate_alibi(
max_seq_len=seqlen_alibi,
num_attention_heads=nheads,
tp_world_size=1,
tp_index=0,
key_padding_mask=None,
device="cuda"
)
# alibi_tensor = alibi_tensor.expand(batch_size, -1, seqlen_q, -1)
alibi_slopes = repeat(alibi_slopes, "nh -> b nh", b=batch_size)
if alibi_tensor.abs().max().item() >= torch.finfo(dtype).max:
pytest.skip()
else:
alibi_tensor, alibi_slopes = None, None
out = flash_attn_with_kvcache(
q,
k_cache,
v_cache,
k,
v,
cos,
sin,
cache_seqlens,
cache_batch_idx,
causal=causal,
window_size=window_size,
rotary_interleaved=rotary_interleaved,
num_splits=num_splits,
alibi_slopes=alibi_slopes
)
# out = flash_attn_with_kvcache(
# q, k_cache, v_cache, cache_seqlens=cache_seqlens, causal=causal, window_size=window_size
# )
# out = flash_attn_with_kvcache(q, k_cache, v_cache, causal=causal, window_size=window_size)
# qk = torch.einsum("bqhd,bkhd->bhqk", q, k_cache_ref)
# m = qk.amax(-1, keepdim=True)
# s_tmp = torch.exp((qk - m) / math.sqrt(d))
# o1 = torch.einsum('bhst,bthd->bshd', s_tmp, v_cache_ref)
# lse_ref = torch.logsumexp(qk / math.sqrt(d), -1)
# probs = torch.softmax(qk, dim=-1)
key_padding_mask = arange < cache_seqlens_expanded + \
(seqlen_new if new_kv else 0)
out_ref, _ = attention_ref(
q_ro,
k_cache_rep,
v_cache_rep,
None,
key_padding_mask,
0.0,
None,
causal=causal,
window_size=window_size,
bias=alibi_tensor
)
out_pt, _ = attention_ref(
q_ro,
k_cache_rep,
v_cache_rep,
None,
key_padding_mask,
0.0,
None,
causal=causal,
window_size=window_size,
upcast=False,
reorder_ops=True,
bias=alibi_tensor
)
print(f"Output max diff: {(out - out_ref).abs().max().item()}")
print(f"Output mean diff: {(out - out_ref).abs().mean().item()}")
print(f"Pytorch max diff: {(out_pt - out_ref).abs().max().item()}")
print(f"Pytorch mean diff: {(out_pt - out_ref).abs().mean().item()}")
# Check that FlashAttention's numerical error is at most twice the numerical error
# of a Pytorch implementation.
if new_kv:
k_cache_select = k_cache if not has_batch_idx else k_cache[cache_batch_idx]
v_cache_select = v_cache if not has_batch_idx else v_cache[cache_batch_idx]
assert torch.allclose(k_cache_select, k_cache_ref,
rtol=1e-3, atol=1e-3)
assert torch.equal(v_cache_select, v_cache_ref)
assert (out - out_ref).abs().max().item() <= 3 * \
(out_pt - out_ref).abs().max().item() + 1e-5
tests/test_flash_attn.py
View file @ 5ab9b366
......@@ -26,6 +26,31 @@ is_sm80 = torch.cuda.get_device_capability("cuda") == (8, 0)
is_sm90 = torch.cuda.get_device_capability("cuda") == (9, 0)
def attn_bias_from_alibi_slopes(
slopes, seqlen_q, seqlen_k, query_padding_mask=None, key_padding_mask=None, causal=False
):
batch, nheads = slopes.shape
device = slopes.device
slopes = rearrange(slopes, "b h -> b h 1 1")
if causal:
return torch.arange(-seqlen_k + 1, 1, device=device, dtype=torch.float32) * slopes
else:
row_idx = rearrange(torch.arange(seqlen_q, device=device, dtype=torch.long), "s -> s 1")
col_idx = torch.arange(seqlen_k, device=device, dtype=torch.long)
sk = (
seqlen_k
if key_padding_mask is None
else rearrange(key_padding_mask.sum(-1), "b -> b 1 1 1")
)
sq = (
seqlen_q
if query_padding_mask is None
else rearrange(query_padding_mask.sum(-1), "b -> b 1 1 1")
)
relative_pos = torch.abs(row_idx + sk - sq - col_idx)
return -slopes * relative_pos.to(dtype=slopes.dtype)
def generate_random_padding_mask(max_seqlen, batch_size, device, mode="random"):
assert mode in ["full", "random", "third"]
if mode == "full":
......@@ -186,6 +211,7 @@ def attention_ref(
v,
query_padding_mask=None,
key_padding_mask=None,
attn_bias=None,
dropout_p=0.0,
dropout_mask=None,
causal=False,
......@@ -200,6 +226,7 @@ def attention_ref(
v: (batch_size, seqlen_k, nheads_k, head_dim)
query_padding_mask: (batch_size, seqlen_q)
key_padding_mask: (batch_size, seqlen_k)
attn_bias: broadcastable to (batch_size, nheads, seqlen_q, seqlen_k)
dropout_p: float
dropout_mask: (batch_size, nheads, seqlen_q, seqlen_k)
causal: whether to apply causal masking
......@@ -238,7 +265,9 @@ def attention_ref(
q.device,
)
scores.masked_fill_(local_mask, float("-inf"))
attention = torch.softmax(scores, dim=-1)
if attn_bias is not None:
scores = scores + attn_bias
attention = torch.softmax(scores, dim=-1).to(v.dtype)
# Some rows might be completely masked out so we fill them with zero instead of NaN
if window_size[0] >= 0 or window_size[1] >= 0:
attention = attention.masked_fill(torch.all(local_mask, dim=-1, keepdim=True), 0.0)
......@@ -264,6 +293,7 @@ def attention_kvpacked_ref(
kv,
query_padding_mask=None,
key_padding_mask=None,
attn_bias=None,
dropout_p=0.0,
dropout_mask=None,
causal=False,
......@@ -277,6 +307,7 @@ def attention_kvpacked_ref(
kv[:, :, 1],
query_padding_mask,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
upcast=upcast,
......@@ -289,6 +320,7 @@ def attention_kvpacked_ref(
def attention_qkvpacked_ref(
qkv,
key_padding_mask=None,
attn_bias=None,
dropout_p=0.0,
dropout_mask=None,
causal=False,
......@@ -302,6 +334,7 @@ def attention_qkvpacked_ref(
qkv[:, :, 2],
key_padding_mask,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
upcast=upcast,
......@@ -436,6 +469,7 @@ def normalize_flash_attn_S(
v,
query_padding_mask=None,
key_padding_mask=None,
attn_bias=None,
is_dropout=False,
causal=False,
window_size=(-1, -1), # -1 means infinite window size
......@@ -445,6 +479,7 @@ def normalize_flash_attn_S(
q: (batch_size, seqlen_q, nheads, head_dim)
k, v: (batch_size, seqlen_k, nheads, head_dim)
key_padding_mask: (batch_size, seqlen_q)
attn_bias: broadcastable to (batch_size, nheads, seqlen_q, seqlen_k)
Output:
softmax_lse: (batch_size, nheads, seqlen_q)
softmax_max: (batch_size, nheads, seqlen_q)
......@@ -467,6 +502,8 @@ def normalize_flash_attn_S(
q.device,
)
scores.masked_fill_(local_mask, float("-inf"))
if attn_bias is not None:
scores = scores + attn_bias.to(dtype=scores.dtype)
_, block_size_n = _get_block_size(scores.device, head_dim, is_dropout, causal)
scores_block = scores.split(block_size_n, dim=-1)
lse_block = torch.stack([torch.logsumexp(s, dim=-1) for s in scores_block], dim=-1)
......@@ -529,6 +566,8 @@ def get_dropout_fraction(
@pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
# @pytest.mark.parametrize("dtype", [torch.float16])
@pytest.mark.parametrize("alibi", [False, True])
# @pytest.mark.parametrize("alibi", [True])
@pytest.mark.parametrize("local", [False, True])
# @pytest.mark.parametrize("local", [True])
@pytest.mark.parametrize("causal", [False, True])
......@@ -538,24 +577,34 @@ def get_dropout_fraction(
# @pytest.mark.parametrize('d', [32, 64, 96, 128])
# @pytest.mark.parametrize("d", [64])
# @pytest.mark.parametrize('seqlen', [128, 256, 384, 512, 768, 1024, 2048])
@pytest.mark.parametrize("seqlen", [97, 128, 200, 256, 257, 384, 512, 768, 1024, 1025, 2048])
# @pytest.mark.parametrize("seqlen", [128])
@pytest.mark.parametrize("seqlen", [97, 128, 200, 384, 768, 1024, 1025, 2048])
# @pytest.mark.parametrize("seqlen", [97])
@pytest.mark.parametrize("dropout_p", [0.0, 0.17])
# @pytest.mark.parametrize("dropout_p", [0.0])
def test_flash_attn_qkvpacked(seqlen, d, dropout_p, causal, local, dtype):
def test_flash_attn_qkvpacked(seqlen, d, dropout_p, causal, local, alibi, dtype):
if seqlen >= 2048 and torch.cuda.get_device_properties("cuda").total_memory <= 16 * 2**30:
pytest.skip() # Reference implementation OOM
device = "cuda"
# set seed
torch.random.manual_seed(0)
batch_size = 13
batch_size = 8
nheads = 9
window_size = (-1, -1) if not local else torch.randint(0, seqlen, (2,))
qkv = torch.randn(
batch_size, seqlen, 3, nheads, d, device=device, dtype=dtype, requires_grad=True
)
if alibi:
alibi_slopes = torch.rand(batch_size, nheads, device=device, dtype=torch.float32) * 0.3
attn_bias = attn_bias_from_alibi_slopes(alibi_slopes, seqlen, seqlen, causal=causal)
else:
alibi_slopes, attn_bias = None, None
out, lse, S_dmask = flash_attn_qkvpacked_func(
qkv, dropout_p, causal=causal, window_size=window_size, return_attn_probs=True
qkv,
dropout_p,
causal=causal,
window_size=window_size,
alibi_slopes=alibi_slopes,
return_attn_probs=True,
)
if dropout_p > 0.0:
S_dmask_converted = convert_flash_attn_S_to_softmax(
......@@ -578,6 +627,7 @@ def test_flash_attn_qkvpacked(seqlen, d, dropout_p, causal, local, dtype):
qkv[:, :, 2],
None,
None,
attn_bias,
dropout_p > 0.0,
causal=causal,
window_size=window_size,
......@@ -590,11 +640,12 @@ def test_flash_attn_qkvpacked(seqlen, d, dropout_p, causal, local, dtype):
dropout_mask = None
out_ref, attn_ref = attention_qkvpacked_ref(
qkv, None, dropout_p, dropout_mask, causal=causal, window_size=window_size
qkv, None, attn_bias, dropout_p, dropout_mask, causal=causal, window_size=window_size
)
out_pt, attn_pt = attention_qkvpacked_ref(
qkv,
None,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -651,7 +702,9 @@ def test_flash_attn_qkvpacked(seqlen, d, dropout_p, causal, local, dtype):
if dropout_p > 0.0:
assert (attn - attn_ref).abs().max().item() <= 2 * (attn_pt - attn_ref).abs().max().item()
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
# With alibi, many of the prob values are 0.0 & -0.0 so dropout_fraction isn't accurate
if not alibi:
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
if d <= MAX_HEADDIM_SM8x or (is_sm80 or is_sm90):
assert (dqkv - dqkv_ref).abs().max().item() <= 2 * (dqkv_pt - dqkv_ref).abs().max().item()
......@@ -659,18 +712,20 @@ def test_flash_attn_qkvpacked(seqlen, d, dropout_p, causal, local, dtype):
@pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
# @pytest.mark.parametrize('dtype', [torch.float16])
@pytest.mark.parametrize("alibi", [False, True])
# @pytest.mark.parametrize("alibi", [True])
@pytest.mark.parametrize("local", [False, True])
# @pytest.mark.parametrize("local", [True])
@pytest.mark.parametrize("causal", [False, True])
# @pytest.mark.parametrize('causal', [False])
@pytest.mark.parametrize("d", [32, 40, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256])
@pytest.mark.parametrize("d", [32, 59, 64, 80, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize("d", [32, 64, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize('d', [64])
@pytest.mark.parametrize("seqlen", [97, 128, 200, 256, 257, 384, 512, 768, 1024, 1025, 2048])
@pytest.mark.parametrize("seqlen", [97, 128, 200, 257, 384, 512, 768, 1025, 2048])
# @pytest.mark.parametrize('seqlen', [128])
@pytest.mark.parametrize("dropout_p", [0.0, 0.17])
# @pytest.mark.parametrize('dropout_p', [0.0])
def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype):
def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, alibi, dtype):
if seqlen >= 2048 and torch.cuda.get_device_properties("cuda").total_memory <= 16 * 2**30:
pytest.skip() # Reference implementation OOM
device = "cuda"
......@@ -685,6 +740,13 @@ def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype)
key_padding_mask = generate_random_padding_mask(seqlen, batch_size, device, mode="random")
# key_padding_mask = generate_random_padding_mask(seqlen, batch_size, device, mode='full')
if alibi:
alibi_slopes = torch.rand(batch_size, nheads, device=device, dtype=torch.float32) * 0.3
attn_bias = attn_bias_from_alibi_slopes(
alibi_slopes, seqlen, seqlen, key_padding_mask, key_padding_mask, causal=causal
)
else:
alibi_slopes, attn_bias = None, None
qkv_unpad, cu_seqlens, max_seqlen, qkv, output_pad_fn, dqkv_pad_fn = generate_qkv(
*qkv.unbind(dim=2), key_padding_mask, key_padding_mask, qkvpacked=True
......@@ -697,6 +759,7 @@ def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype)
dropout_p,
causal=causal,
window_size=window_size,
alibi_slopes=alibi_slopes,
return_attn_probs=True,
)
out = output_pad_fn(out_unpad)
......@@ -721,6 +784,7 @@ def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype)
qkv[:, :, 2],
key_padding_mask,
key_padding_mask,
attn_bias,
dropout_p > 0.0,
causal=causal,
window_size=window_size,
......@@ -733,11 +797,18 @@ def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype)
dropout_mask = None
out_ref, attn_ref = attention_qkvpacked_ref(
qkv, key_padding_mask, dropout_p, dropout_mask, causal=causal, window_size=window_size
qkv,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
window_size=window_size,
)
out_pt, attn_pt = attention_qkvpacked_ref(
qkv,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -774,7 +845,9 @@ def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype)
if dropout_p > 0.0:
assert (attn - attn_ref).abs().max().item() <= 2 * (attn_pt - attn_ref).abs().max().item()
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
# With alibi, many of the prob values are 0.0 & -0.0 so dropout_fraction isn't accurate
if not alibi:
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
if d <= MAX_HEADDIM_SM8x or (is_sm80 or is_sm90):
assert (dqkv - dqkv_ref).abs().max().item() <= 2 * (dqkv_pt - dqkv_ref).abs().max().item()
......@@ -786,11 +859,13 @@ def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype)
# @pytest.mark.parametrize("dtype", [torch.bfloat16])
@pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"])
# @pytest.mark.parametrize("mha_type", ["mha"])
@pytest.mark.parametrize("alibi", [False, True])
# @pytest.mark.parametrize("alibi", [True])
@pytest.mark.parametrize("local", [False, True])
# @pytest.mark.parametrize("local", [True])
@pytest.mark.parametrize("causal", [False, True])
# @pytest.mark.parametrize("causal", [True])
@pytest.mark.parametrize("d", [32, 40, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256])
@pytest.mark.parametrize("d", [32, 40, 59, 64, 96, 111, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize("d", [32, 64, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize('d', [32, 40, 64, 80, 96, 128, 160, 192])
# @pytest.mark.parametrize('d', [32, 64, 96, 128, 160, 192])
......@@ -815,7 +890,7 @@ def test_flash_attn_varlen_qkvpacked(seqlen, d, dropout_p, causal, local, dtype)
@pytest.mark.parametrize("dropout_p", [0.0, 0.17])
# @pytest.mark.parametrize("dropout_p", [0.17])
def test_flash_attn_output(
seqlen_q, seqlen_k, d, dropout_p, causal, local, mha_type, dtype, kvpacked
seqlen_q, seqlen_k, d, dropout_p, causal, local, alibi, mha_type, dtype, kvpacked
):
if (
max(seqlen_q, seqlen_k) >= 2048
......@@ -825,7 +900,7 @@ def test_flash_attn_output(
device = "cuda"
# set seed
torch.random.manual_seed(0)
batch_size = 13
batch_size = 8
nheads = 9
nheads_k = nheads if mha_type == "mha" else (1 if mha_type == "mqa" else 3)
assert nheads % nheads_k == 0
......@@ -842,14 +917,32 @@ def test_flash_attn_output(
v = torch.randn(
batch_size, seqlen_k, nheads_k, d, device=device, dtype=dtype, requires_grad=True
)
if alibi:
alibi_slopes = torch.rand(batch_size, nheads, device=device, dtype=torch.float32) * 0.3
attn_bias = attn_bias_from_alibi_slopes(alibi_slopes, seqlen_q, seqlen_k, causal=causal)
else:
alibi_slopes, attn_bias = None, None
if kvpacked:
out, lse, S_dmask = flash_attn_kvpacked_func(
q, kv, dropout_p, causal=causal, window_size=window_size, return_attn_probs=True
q,
kv,
dropout_p,
causal=causal,
window_size=window_size,
alibi_slopes=alibi_slopes,
return_attn_probs=True,
)
else:
out, lse, S_dmask = flash_attn_func(
q, k, v, dropout_p, causal=causal, window_size=window_size, return_attn_probs=True
q,
k,
v,
dropout_p,
causal=causal,
window_size=window_size,
alibi_slopes=alibi_slopes,
return_attn_probs=True,
)
if dropout_p > 0.0:
S_dmask_converted = convert_flash_attn_S_to_softmax(
......@@ -878,6 +971,7 @@ def test_flash_attn_output(
v_rep,
None,
None,
attn_bias,
dropout_p > 0.0,
causal=causal,
window_size=window_size,
......@@ -895,6 +989,7 @@ def test_flash_attn_output(
kv,
None,
None,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -905,6 +1000,7 @@ def test_flash_attn_output(
kv,
None,
None,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -919,6 +1015,7 @@ def test_flash_attn_output(
v,
None,
None,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -930,6 +1027,7 @@ def test_flash_attn_output(
v,
None,
None,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -1000,7 +1098,9 @@ def test_flash_attn_output(
if dropout_p > 0.0:
assert (attn - attn_ref).abs().max().item() <= 2 * (attn_pt - attn_ref).abs().max().item()
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
# With alibi, many of the prob values are 0.0 & -0.0 so dropout_fraction isn't accurate
if not alibi:
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
if d <= MAX_HEADDIM_SM8x or (is_sm80 or is_sm90):
assert (dq - dq_ref).abs().max().item() <= 2 * (dq_pt - dq_ref).abs().max().item()
......@@ -1014,11 +1114,13 @@ def test_flash_attn_output(
# @pytest.mark.parametrize('dtype', [torch.float16])
@pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"])
# @pytest.mark.parametrize('mha_type', ["mqa"])
@pytest.mark.parametrize("alibi", [False, True])
# @pytest.mark.parametrize("alibi", [True])
@pytest.mark.parametrize("local", [False, True])
# @pytest.mark.parametrize("local", [True])
@pytest.mark.parametrize("causal", [False, True])
# @pytest.mark.parametrize('causal', [True])
@pytest.mark.parametrize("d", [32, 40, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256])
@pytest.mark.parametrize("d", [32, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize("d", [32, 64, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize('d', [64])
@pytest.mark.parametrize(
......@@ -1041,7 +1143,7 @@ def test_flash_attn_output(
@pytest.mark.parametrize("dropout_p", [0.0, 0.17])
# @pytest.mark.parametrize('dropout_p', [0.0])
def test_flash_attn_varlen_output(
seqlen_q, seqlen_k, d, dropout_p, causal, local, mha_type, dtype, kvpacked
seqlen_q, seqlen_k, d, dropout_p, causal, local, alibi, mha_type, dtype, kvpacked
):
if (
max(seqlen_q, seqlen_k) >= 2048
......@@ -1051,7 +1153,7 @@ def test_flash_attn_varlen_output(
device = "cuda"
# set seed
torch.random.manual_seed(0)
batch_size = 13
batch_size = 8
nheads = 9
nheads_k = nheads if mha_type == "mha" else (1 if mha_type == "mqa" else 3)
assert nheads % nheads_k == 0
......@@ -1072,6 +1174,13 @@ def test_flash_attn_varlen_output(
query_padding_mask = generate_random_padding_mask(seqlen_q, batch_size, device, mode="random")
key_padding_mask = generate_random_padding_mask(seqlen_k, batch_size, device, mode="random")
# key_padding_mask = generate_random_padding_mask(seqlen_k, batch_size, device, mode='full')
if alibi:
alibi_slopes = torch.rand(batch_size, nheads, device=device, dtype=torch.float32) * 0.3
attn_bias = attn_bias_from_alibi_slopes(
alibi_slopes, seqlen_q, seqlen_k, query_padding_mask, key_padding_mask, causal=causal
)
else:
alibi_slopes, attn_bias = None, None
if kvpacked:
(
......@@ -1095,9 +1204,10 @@ def test_flash_attn_varlen_output(
max_seqlen_q,
max_seqlen_k,
dropout_p,
return_attn_probs=True,
causal=causal,
window_size=window_size,
alibi_slopes=alibi_slopes,
return_attn_probs=True,
)
else:
(
......@@ -1124,9 +1234,10 @@ def test_flash_attn_varlen_output(
max_seqlen_q,
max_seqlen_k,
dropout_p,
return_attn_probs=True,
causal=causal,
window_size=window_size,
alibi_slopes=alibi_slopes,
return_attn_probs=True,
)
out = output_pad_fn(out_unpad)
if dropout_p > 0.0:
......@@ -1156,6 +1267,7 @@ def test_flash_attn_varlen_output(
v_rep,
query_padding_mask,
key_padding_mask,
attn_bias,
dropout_p > 0.0,
causal=causal,
window_size=window_size,
......@@ -1177,6 +1289,7 @@ def test_flash_attn_varlen_output(
kv,
query_padding_mask,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -1187,6 +1300,7 @@ def test_flash_attn_varlen_output(
kv,
query_padding_mask,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -1201,6 +1315,7 @@ def test_flash_attn_varlen_output(
v,
query_padding_mask,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -1212,6 +1327,7 @@ def test_flash_attn_varlen_output(
v,
query_padding_mask,
key_padding_mask,
attn_bias,
dropout_p,
dropout_mask,
causal=causal,
......@@ -1284,12 +1400,14 @@ def test_flash_attn_varlen_output(
if dropout_p > 0.0:
assert (attn - attn_ref).abs().max().item() <= 2 * (attn_pt - attn_ref).abs().max().item()
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
# With alibi, many of the prob values are 0.0 & -0.0 so dropout_fraction isn't accurate
if not alibi:
assert abs(dropout_fraction - dropout_p) <= (0.01 if not local else 0.025)
if d <= MAX_HEADDIM_SM8x or (is_sm80 or is_sm90):
assert (dq - dq_ref).abs().max().item() <= 2 * (dq_pt - dq_ref).abs().max().item()
assert (dk - dk_ref).abs().max().item() <= 2 * (dk_pt - dk_ref).abs().max().item()
assert (dv - dv_ref).abs().max().item() <= 2 * (dv_pt - dv_ref).abs().max().item()
assert (dq - dq_ref).abs().max().item() <= 3 * (dq_pt - dq_ref).abs().max().item()
assert (dk - dk_ref).abs().max().item() <= 3 * (dk_pt - dk_ref).abs().max().item()
assert (dv - dv_ref).abs().max().item() <= 3 * (dv_pt - dv_ref).abs().max().item()
@pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
......@@ -1332,7 +1450,7 @@ def test_flash_attn_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtype):
causal = True
# set seed
torch.random.manual_seed(0)
batch_size = 13
batch_size = 8
nheads = 9
window_size = (-1, -1) if not local else torch.randint(0, seqlen_k, (2,))
q = torch.randn(batch_size, seqlen_q, nheads, d, device=device, dtype=dtype, requires_grad=True)
......@@ -1340,7 +1458,7 @@ def test_flash_attn_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtype):
v = torch.randn(batch_size, seqlen_k, nheads, d, device=device, dtype=dtype, requires_grad=True)
out = flash_attn_func(q, k, v, 0.0, causal=causal, window_size=window_size)
out_ref, attn_ref = attention_ref(
q, k, v, None, None, 0.0, None, causal=causal, window_size=window_size
q, k, v, None, None, None, 0.0, None, causal=causal, window_size=window_size
)
out_pt, attn_pt = attention_ref(
q,
......@@ -1348,6 +1466,7 @@ def test_flash_attn_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtype):
v,
None,
None,
None,
0.0,
None,
causal=causal,
......@@ -1442,7 +1561,7 @@ def test_flash_attn_varlen_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtyp
causal = True
# set seed
torch.random.manual_seed(0)
batch_size = 13
batch_size = 8
nheads = 9
window_size = (-1, -1) if not local else torch.randint(0, seqlen_k, (2,))
q = torch.randn(batch_size, seqlen_q, nheads, d, device=device, dtype=dtype, requires_grad=True)
......@@ -1484,6 +1603,7 @@ def test_flash_attn_varlen_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtyp
v,
query_padding_mask,
key_padding_mask,
None,
0.0,
None,
causal=causal,
......@@ -1495,6 +1615,7 @@ def test_flash_attn_varlen_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtyp
v,
query_padding_mask,
key_padding_mask,
None,
0.0,
None,
causal=causal,
......@@ -1554,8 +1675,10 @@ def test_flash_attn_varlen_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtyp
@pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
# @pytest.mark.parametrize("dtype", [torch.float16])
@pytest.mark.parametrize("alibi", [False, True])
# @pytest.mark.parametrize("alibi", [True])
@pytest.mark.parametrize("local", [False, True])
# @pytest.mark.parametrize("local", [True])
# @pytest.mark.parametrize("local", [False])
@pytest.mark.parametrize("causal", [False, True])
# @pytest.mark.parametrize("causal", [True])
@pytest.mark.parametrize("d", [32, 40, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256])
......@@ -1581,7 +1704,7 @@ def test_flash_attn_varlen_causal(seqlen_q, seqlen_k, swap_sq_sk, d, local, dtyp
],
)
# @pytest.mark.parametrize('seqlen_q,seqlen_k', [(256, 128)])
def test_flash_attn_splitkv(seqlen_q, seqlen_k, swap_sq_sk, d, causal, local, dtype):
def test_flash_attn_splitkv(seqlen_q, seqlen_k, swap_sq_sk, d, causal, local, alibi, dtype):
if swap_sq_sk:
seqlen_q, seqlen_k = seqlen_k, seqlen_q
device = "cuda"
......@@ -1593,11 +1716,23 @@ def test_flash_attn_splitkv(seqlen_q, seqlen_k, swap_sq_sk, d, causal, local, dt
q = torch.randn(batch_size, seqlen_q, nheads, d, device=device, dtype=dtype, requires_grad=True)
k = torch.randn(batch_size, seqlen_k, nheads, d, device=device, dtype=dtype, requires_grad=True)
v = torch.randn(batch_size, seqlen_k, nheads, d, device=device, dtype=dtype, requires_grad=True)
if alibi:
alibi_slopes = torch.rand(batch_size, nheads, device=device, dtype=torch.float32) * 0.3
attn_bias = attn_bias_from_alibi_slopes(alibi_slopes, seqlen_q, seqlen_k, causal=causal)
else:
alibi_slopes, attn_bias = None, None
out, lse, _ = flash_attn_func(
q, k, v, 0.0, causal=causal, window_size=window_size, return_attn_probs=True
q,
k,
v,
0.0,
causal=causal,
window_size=window_size,
alibi_slopes=alibi_slopes,
return_attn_probs=True,
)
out_ref, attn_ref = attention_ref(
q, k, v, None, None, 0.0, None, causal=causal, window_size=window_size
q, k, v, None, None, attn_bias, 0.0, None, causal=causal, window_size=window_size
)
out_pt, attn_pt = attention_ref(
q,
......@@ -1605,6 +1740,7 @@ def test_flash_attn_splitkv(seqlen_q, seqlen_k, swap_sq_sk, d, causal, local, dt
v,
None,
None,
attn_bias,
0.0,
None,
causal=causal,
......@@ -1653,24 +1789,27 @@ def test_flash_attn_splitkv(seqlen_q, seqlen_k, swap_sq_sk, d, causal, local, dt
# of a Pytorch implementation.
assert (out - out_ref).abs().max().item() <= 2 * (out_pt - out_ref).abs().max().item() + 1e-5
mult = 2 if not alibi else 8
if d <= MAX_HEADDIM_SM8x or (is_sm80 or is_sm90):
assert (dq - dq_ref).abs().max().item() <= 2 * (dq_pt - dq_ref).abs().max().item() + 2e-4
assert (dk - dk_ref).abs().max().item() <= 2 * (dk_pt - dk_ref).abs().max().item() + 2e-4
assert (dv - dv_ref).abs().max().item() <= 2 * (dv_pt - dv_ref).abs().max().item() + 2e-4
assert (dq - dq_ref).abs().max().item() <= mult * (dq_pt - dq_ref).abs().max().item() + 2e-4
assert (dk - dk_ref).abs().max().item() <= mult * (dk_pt - dk_ref).abs().max().item() + 2e-4
assert (dv - dv_ref).abs().max().item() <= mult * (dv_pt - dv_ref).abs().max().item() + 2e-4
@pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
# @pytest.mark.parametrize("dtype", [torch.float16])
# @pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
@pytest.mark.parametrize("dtype", [torch.float16])
@pytest.mark.parametrize("num_splits", [1, 0])
# @pytest.mark.parametrize("num_splits", [0])
# @pytest.mark.parametrize("num_splits", [1])
@pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"])
# @pytest.mark.parametrize("mha_type", ["mha"])
@pytest.mark.parametrize("new_kv", [False, True])
# @pytest.mark.parametrize("new_kv", [True])
# @pytest.mark.parametrize("new_kv", [False])
@pytest.mark.parametrize("alibi", [False, True])
# @pytest.mark.parametrize("alibi", [True])
@pytest.mark.parametrize("local", [False, True])
# @pytest.mark.parametrize("local", [False])
@pytest.mark.parametrize("causal", [False, True])
# @pytest.mark.parametrize("causal", [True])
# @pytest.mark.parametrize("causal", [False])
@pytest.mark.parametrize("seqlen_new_eq_seqlen_q", [True, False])
# @pytest.mark.parametrize("seqlen_new_eq_seqlen_q", [True])
@pytest.mark.parametrize("rotary_interleaved", [False, True])
......@@ -1678,7 +1817,7 @@ def test_flash_attn_splitkv(seqlen_q, seqlen_k, swap_sq_sk, d, causal, local, dt
@pytest.mark.parametrize("rotary_fraction", [0.0, 0.5, 1.0])
# @pytest.mark.parametrize("rotary_fraction", [0.0])
@pytest.mark.parametrize("has_batch_idx", [False, True])
# @pytest.mark.parametrize("has_batch_idx", [True])
# @pytest.mark.parametrize("has_batch_idx", [False])
@pytest.mark.parametrize("d", [32, 59, 64, 80, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize('d', [32, 64, 96, 128, 160, 192, 224, 256])
# @pytest.mark.parametrize('d', [32, 40, 64, 80, 96, 128, 160, 192])
......@@ -1711,6 +1850,7 @@ def test_flash_attn_kvcache(
seqlen_new_eq_seqlen_q,
causal,
local,
alibi,
new_kv,
mha_type,
num_splits,
......@@ -1750,10 +1890,22 @@ def test_flash_attn_kvcache(
dtype=torch.int32,
device=device,
)
arange = rearrange(torch.arange(seqlen_k, device=device), "s -> 1 s")
cache_seqlens_expanded = rearrange(cache_seqlens, "b -> b 1")
key_padding_mask = arange < cache_seqlens_expanded + (seqlen_new if new_kv else 0)
if has_batch_idx:
cache_batch_idx = torch.randperm(batch_size_cache, dtype=torch.int32, device=device)[:batch_size]
cache_batch_idx = torch.randperm(batch_size_cache, dtype=torch.int32, device=device)[
:batch_size
]
else:
cache_batch_idx = None
if alibi:
alibi_slopes = torch.rand(batch_size, nheads, device=device, dtype=torch.float32) * 0.3
attn_bias = attn_bias_from_alibi_slopes(
alibi_slopes, seqlen_q, seqlen_k, None, key_padding_mask, causal=causal
)
else:
alibi_slopes, attn_bias = None, None
# cache_seqlens = torch.tensor([64], dtype=torch.int32, device=device)
if rotary_dim > 0:
angle = torch.rand(seqlen_k, rotary_dim // 2, device=device) * 2 * math.pi
......@@ -1785,8 +1937,6 @@ def test_flash_attn_kvcache(
# k_cache[:, 64:] = -1
k_cache_ref = (k_cache if not has_batch_idx else k_cache[cache_batch_idx]).clone()
v_cache_ref = (v_cache if not has_batch_idx else v_cache[cache_batch_idx]).clone()
arange = rearrange(torch.arange(seqlen_k, device=device), "s -> 1 s")
cache_seqlens_expanded = rearrange(cache_seqlens, "b -> b 1")
if new_kv:
update_mask = torch.logical_and(
cache_seqlens_expanded <= arange, arange < cache_seqlens_expanded + seqlen_new
......@@ -1808,6 +1958,7 @@ def test_flash_attn_kvcache(
causal=causal,
window_size=window_size,
rotary_interleaved=rotary_interleaved,
alibi_slopes=alibi_slopes,
num_splits=num_splits,
)
# out = flash_attn_with_kvcache(
......@@ -1820,13 +1971,13 @@ def test_flash_attn_kvcache(
# o1 = torch.einsum('bhst,bthd->bshd', s_tmp, v_cache_ref)
# lse_ref = torch.logsumexp(qk / math.sqrt(d), -1)
# probs = torch.softmax(qk, dim=-1)
key_padding_mask = arange < cache_seqlens_expanded + (seqlen_new if new_kv else 0)
out_ref, _ = attention_ref(
q_ro,
k_cache_rep,
v_cache_rep,
None,
key_padding_mask,
attn_bias,
0.0,
None,
causal=causal,
......@@ -1838,6 +1989,7 @@ def test_flash_attn_kvcache(
v_cache_rep,
None,
key_padding_mask,
attn_bias,
0.0,
None,
causal=causal,
......@@ -1857,7 +2009,8 @@ def test_flash_attn_kvcache(
v_cache_select = v_cache if not has_batch_idx else v_cache[cache_batch_idx]
assert torch.allclose(k_cache_select, k_cache_ref, rtol=1e-3, atol=1e-3)
assert torch.equal(v_cache_select, v_cache_ref)
assert (out - out_ref).abs().max().item() <= 3 * (out_pt - out_ref).abs().max().item() + 1e-5
mult = 3 if not alibi else 5
assert (out - out_ref).abs().max().item() <= mult * (out_pt - out_ref).abs().max().item() + 1e-5
# @pytest.mark.parametrize("dtype", ([torch.float16] if is_sm75 else [torch.float16, torch.bfloat16]))
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment