root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 03:14:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

ggml : fix YARN + add tests + add asserts (#7617)

Browse Source 
* tests : add rope tests

ggml-ci

* ggml : fixes (hopefully)

ggml-ci

* tests : add non-cont tests

ggml-ci

* cuda : add asserts for rope/norm + fix DS2

ggml-ci

* ggml : assert contiguousness

* tests : reduce RoPE tests

ggml-ci
This commit is contained in:
Georgi Gerganov 2024-05-29 20:17:31 +03:00 committed by GitHub
parent cce3dcffc5
commit fb76ec31a9
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
12 changed files with 167 additions and 104 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
ggml-cuda.cu
View File

@ -1870,7 +1870,7 @@ static void ggml_cuda_mul_mat_batched_cublas(ggml_backend_cuda_context & ctx, co
} }
} }
#else #else
if (r2 == 1 && r3 == 1 && src0->nb[2]*src0->ne[2] == src0->nb[3] && src1->nb[2]*src1->ne[2] == src1->nb[3]) { if (r2 == 1 && r3 == 1 && ggml_is_contiguous_2(src0) && ggml_is_contiguous_2(src1)) {
// there is no broadcast and src0, src1 are contiguous across dims 2, 3 // there is no broadcast and src0, src1 are contiguous across dims 2, 3
// use cublasGemmStridedBatchedEx // use cublasGemmStridedBatchedEx
CUBLAS_CHECK( CUBLAS_CHECK(
@ -2886,7 +2886,9 @@ GGML_CALL static bool ggml_backend_cuda_supports_op(ggml_backend_t backend, cons
case GGML_OP_CONT: case GGML_OP_CONT:
case GGML_OP_DIAG_MASK_INF: case GGML_OP_DIAG_MASK_INF:
case GGML_OP_SOFT_MAX: case GGML_OP_SOFT_MAX:
return true;
case GGML_OP_ROPE: case GGML_OP_ROPE:
return ggml_is_contiguous(op->src[0]);
case GGML_OP_IM2COL: case GGML_OP_IM2COL:
case GGML_OP_POOL_2D: case GGML_OP_POOL_2D:
case GGML_OP_SUM_ROWS: case GGML_OP_SUM_ROWS:

6
ggml-cuda/norm.cu
View File

@ -170,6 +170,8 @@ void ggml_cuda_op_norm(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
float * dst_d = (float *)dst->data; float * dst_d = (float *)dst->data;
cudaStream_t stream = ctx.stream(); cudaStream_t stream = ctx.stream();
GGML_ASSERT(ggml_is_contiguous(src0));
GGML_ASSERT(src0->type == GGML_TYPE_F32); GGML_ASSERT(src0->type == GGML_TYPE_F32);
GGML_ASSERT( dst->type == GGML_TYPE_F32); GGML_ASSERT( dst->type == GGML_TYPE_F32);
@ -188,6 +190,8 @@ void ggml_cuda_op_group_norm(ggml_backend_cuda_context & ctx, ggml_tensor * dst)
float * dst_d = (float *)dst->data; float * dst_d = (float *)dst->data;
cudaStream_t stream = ctx.stream(); cudaStream_t stream = ctx.stream();
GGML_ASSERT(ggml_is_contiguous(src0));
GGML_ASSERT(src0->type == GGML_TYPE_F32); GGML_ASSERT(src0->type == GGML_TYPE_F32);
GGML_ASSERT( dst->type == GGML_TYPE_F32); GGML_ASSERT( dst->type == GGML_TYPE_F32);
@ -202,6 +206,8 @@ void ggml_cuda_op_rms_norm(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
float * dst_d = (float *)dst->data; float * dst_d = (float *)dst->data;
cudaStream_t stream = ctx.stream(); cudaStream_t stream = ctx.stream();
GGML_ASSERT(ggml_is_contiguous(src0));
GGML_ASSERT(src0->type == GGML_TYPE_F32); GGML_ASSERT(src0->type == GGML_TYPE_F32);
GGML_ASSERT( dst->type == GGML_TYPE_F32); GGML_ASSERT( dst->type == GGML_TYPE_F32);

18
ggml-cuda/rope.cu
View File

@ -61,7 +61,7 @@ static __global__ void rope(
template<typename T, bool has_pos, bool has_freq_facs> template<typename T, bool has_pos, bool has_freq_facs>
static __global__ void rope_neox( static __global__ void rope_neox(
const T * x, T * dst, int ncols, int n_dims, const int32_t * pos, float freq_scale, int p_delta_rows, const T * x, T * dst, int ncols, int n_dims, const int32_t * pos, float freq_scale, int p_delta_rows,
float ext_factor, float attn_factor, rope_corr_dims corr_dims, float theta_scale, float inv_ndims, const float * freq_factors float ext_factor, float attn_factor, rope_corr_dims corr_dims, float theta_scale, const float * freq_factors
) { ) {
const int col = 2*(blockDim.y*blockIdx.y + threadIdx.y); const int col = 2*(blockDim.y*blockIdx.y + threadIdx.y);
@ -85,15 +85,13 @@ static __global__ void rope_neox(
const int i = row*ncols + ib*n_dims + ic/2; const int i = row*ncols + ib*n_dims + ic/2;
const int i2 = row/p_delta_rows; const int i2 = row/p_delta_rows;
float cur_rot = inv_ndims * ic - ib;
const int p = has_pos ? pos[i2] : 0; const int p = has_pos ? pos[i2] : 0;
const float freq_factor = has_freq_facs ? freq_factors[ic/2] : 1.0f; const float freq_factor = has_freq_facs ? freq_factors[ic/2] : 1.0f;
const float theta_base = p*freq_scale*powf(theta_scale, col/2.0f)/freq_factor; const float theta_base = p*powf(theta_scale, col/2.0f)/freq_factor;
float cos_theta, sin_theta; float cos_theta, sin_theta;
rope_yarn(theta_base, freq_scale, corr_dims, cur_rot, ext_factor, attn_factor, &cos_theta, &sin_theta); rope_yarn(theta_base, freq_scale, corr_dims, ic, ext_factor, attn_factor, &cos_theta, &sin_theta);
const float x0 = x[i + 0]; const float x0 = x[i + 0];
const float x1 = x[i + n_dims/2]; const float x1 = x[i + n_dims/2];
@ -174,30 +172,29 @@ static void rope_neox_cuda(
const dim3 block_nums(nrows, num_blocks_x, 1); const dim3 block_nums(nrows, num_blocks_x, 1);
const float theta_scale = powf(freq_base, -2.0f/n_dims); const float theta_scale = powf(freq_base, -2.0f/n_dims);
const float inv_ndims = -1.0f / n_dims;
if (pos == nullptr) { if (pos == nullptr) {
if (freq_factors == nullptr) { if (freq_factors == nullptr) {
rope_neox<T, false, false><<<block_nums, block_dims, 0, stream>>>( rope_neox<T, false, false><<<block_nums, block_dims, 0, stream>>>(
x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims,
theta_scale, inv_ndims, freq_factors theta_scale, freq_factors
); );
} else { } else {
rope_neox<T, false, true><<<block_nums, block_dims, 0, stream>>>( rope_neox<T, false, true><<<block_nums, block_dims, 0, stream>>>(
x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims,
theta_scale, inv_ndims, freq_factors theta_scale, freq_factors
); );
} }
} else { } else {
if (freq_factors == nullptr) { if (freq_factors == nullptr) {
rope_neox<T, true, false><<<block_nums, block_dims, 0, stream>>>( rope_neox<T, true, false><<<block_nums, block_dims, 0, stream>>>(
x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims,
theta_scale, inv_ndims, freq_factors theta_scale, freq_factors
); );
} else { } else {
rope_neox<T, true, true><<<block_nums, block_dims, 0, stream>>>( rope_neox<T, true, true><<<block_nums, block_dims, 0, stream>>>(
x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, x, dst, ncols, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims,
theta_scale, inv_ndims, freq_factors theta_scale, freq_factors
); );
} }
} }
@ -254,6 +251,7 @@ void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
float * dst_d = (float *)dst->data; float * dst_d = (float *)dst->data;
cudaStream_t stream = ctx.stream(); cudaStream_t stream = ctx.stream();
GGML_ASSERT(ggml_is_contiguous(src0));
GGML_ASSERT(src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16); GGML_ASSERT(src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16);
GGML_ASSERT( dst->type == GGML_TYPE_F32 || dst->type == GGML_TYPE_F16); GGML_ASSERT( dst->type == GGML_TYPE_F32 || dst->type == GGML_TYPE_F16);
GGML_ASSERT(src0->type == dst->type); GGML_ASSERT(src0->type == dst->type);

4
ggml-kompute.cpp
View File

@ -1597,7 +1597,9 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml
{ {
GGML_ASSERT(ne00 == ne10); GGML_ASSERT(ne00 == ne10);
// TODO: assert that dim2 and dim3 are contiguous ggml_is_contiguous_2(src0);
ggml_is_contiguous_2(src1);
GGML_ASSERT(ne12 % ne02 == 0); GGML_ASSERT(ne12 % ne02 == 0);
GGML_ASSERT(ne13 % ne03 == 0); GGML_ASSERT(ne13 % ne03 == 0);

8
ggml-metal.m
View File

@ -1519,7 +1519,9 @@ static enum ggml_status ggml_metal_graph_compute(
{ {
GGML_ASSERT(ne00 == ne10); GGML_ASSERT(ne00 == ne10);
// TODO: assert that dim2 and dim3 are contiguous ggml_is_contiguous_2(src0);
ggml_is_contiguous_2(src1);
GGML_ASSERT(ne12 % ne02 == 0); GGML_ASSERT(ne12 % ne02 == 0);
GGML_ASSERT(ne13 % ne03 == 0); GGML_ASSERT(ne13 % ne03 == 0);
@ -2187,6 +2189,7 @@ static enum ggml_status ggml_metal_graph_compute(
case GGML_OP_RMS_NORM: case GGML_OP_RMS_NORM:
{ {
GGML_ASSERT(ne00 % 4 == 0); GGML_ASSERT(ne00 % 4 == 0);
GGML_ASSERT(ggml_is_contiguous_1(src0));
float eps; float eps;
memcpy(&eps, dst->op_params, sizeof(float)); memcpy(&eps, dst->op_params, sizeof(float));
@ -2214,6 +2217,7 @@ static enum ggml_status ggml_metal_graph_compute(
case GGML_OP_GROUP_NORM: case GGML_OP_GROUP_NORM:
{ {
GGML_ASSERT(ne00 % 4 == 0); GGML_ASSERT(ne00 % 4 == 0);
GGML_ASSERT(ggml_is_contiguous(src0));
//float eps; //float eps;
//memcpy(&eps, dst->op_params, sizeof(float)); //memcpy(&eps, dst->op_params, sizeof(float));
@ -2247,6 +2251,8 @@ static enum ggml_status ggml_metal_graph_compute(
} break; } break;
case GGML_OP_NORM: case GGML_OP_NORM:
{ {
GGML_ASSERT(ggml_is_contiguous_1(src0));
float eps; float eps;
memcpy(&eps, dst->op_params, sizeof(float)); memcpy(&eps, dst->op_params, sizeof(float));

16
ggml-metal.metal
View File

@ -1767,13 +1767,13 @@ kernel void kernel_rope(
const int64_t p = pos[i2]; const int64_t p = pos[i2];
const float theta_0 = (float)p; const float theta_base = (float)p;
const float inv_ndims = -1.f/n_dims; const float inv_ndims = -1.f/n_dims;
if (!is_neox) { if (!is_neox) {
for (int64_t i0 = 2*tiitg; i0 < ne0; i0 += 2*tptg.x) { for (int64_t i0 = 2*tiitg; i0 < ne0; i0 += 2*tptg.x) {
const float theta = theta_base * pow(freq_base, inv_ndims*i0);
const float theta = theta_0 * pow(freq_base, inv_ndims*i0);
float cos_theta, sin_theta; float cos_theta, sin_theta;
rope_yarn(theta, freq_scale, corr_dims, i0, ext_factor, attn_factor, &cos_theta, &sin_theta); rope_yarn(theta, freq_scale, corr_dims, i0, ext_factor, attn_factor, &cos_theta, &sin_theta);
@ -1789,18 +1789,14 @@ kernel void kernel_rope(
} else { } else {
for (int64_t ic = 2*tiitg; ic < ne0; ic += 2*tptg.x) { for (int64_t ic = 2*tiitg; ic < ne0; ic += 2*tptg.x) {
if (ic < n_dims) { if (ic < n_dims) {
const int64_t ib = 0; const int64_t i0 = ic/2;
// simplified from `(ib * n_dims + ic) * inv_ndims` const float freq_factor = src2 != src0 ? src2[i0] : 1.0f;
const float cur_rot = inv_ndims*ic - ib;
const float freq_factor = src2 != src0 ? src2[ic/2] : 1.0f;
const float theta = theta_0 * pow(freq_base, cur_rot) / freq_factor; const float theta = theta_base * pow(freq_base, inv_ndims*ic);
float cos_theta, sin_theta; float cos_theta, sin_theta;
rope_yarn(theta, freq_scale, corr_dims, cur_rot, ext_factor, attn_factor, &cos_theta, &sin_theta); rope_yarn(theta/freq_factor, freq_scale, corr_dims, ic, ext_factor, attn_factor, &cos_theta, &sin_theta);
const int64_t i0 = ib*n_dims + ic/2;
device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);

2
ggml-sycl.cpp
View File

@ -15183,7 +15183,7 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0,
const int64_t r2 = ne12/ne02; const int64_t r2 = ne12/ne02;
const int64_t r3 = ne13/ne03; const int64_t r3 = ne13/ne03;
if (r2 == 1 && r3 == 1 && src0->nb[2]*src0->ne[2] == src0->nb[3] && src1->nb[2]*src1->ne[2] == src1->nb[3]) { if (r2 == 1 && r3 == 1 && ggml_is_contiguous_2(src0) && ggml_is_contiguous_2(src1)) {
// there is no broadcast and src0, src1 are contiguous across dims 2, 3 // there is no broadcast and src0, src1 are contiguous across dims 2, 3
SYCL_CHECK(CHECK_TRY_ERROR(dpct::gemm_batch( SYCL_CHECK(CHECK_TRY_ERROR(dpct::gemm_batch(
*g_sycl_handles[g_main_device], oneapi::mkl::transpose::trans, *g_sycl_handles[g_main_device], oneapi::mkl::transpose::trans,

70
ggml.c
View File

@ -3221,7 +3221,11 @@ GGML_CALL bool ggml_is_contiguous(const struct ggml_tensor * tensor) {
tensor->nb[3] == tensor->nb[2]*tensor->ne[2]; tensor->nb[3] == tensor->nb[2]*tensor->ne[2];
} }
static inline bool ggml_is_contiguous_except_dim_1(const struct ggml_tensor * tensor) { GGML_CALL bool ggml_is_contiguous_0(const struct ggml_tensor * tensor) {
return ggml_is_contiguous(tensor);
}
GGML_CALL bool ggml_is_contiguous_1(const struct ggml_tensor * tensor) {
static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function"); static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function");
return return
@ -3230,6 +3234,14 @@ static inline bool ggml_is_contiguous_except_dim_1(const struct ggml_tensor * te
tensor->nb[3] == tensor->nb[2]*tensor->ne[2]; tensor->nb[3] == tensor->nb[2]*tensor->ne[2];
} }
GGML_CALL bool ggml_is_contiguous_2(const struct ggml_tensor * tensor) {
static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function");
return
tensor->nb[0] == ggml_type_size(tensor->type) &&
tensor->nb[3] == tensor->nb[2]*tensor->ne[2];
}
GGML_CALL bool ggml_is_permuted(const struct ggml_tensor * tensor) { GGML_CALL bool ggml_is_permuted(const struct ggml_tensor * tensor) {
static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function"); static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function");
@ -11420,8 +11432,8 @@ static void ggml_compute_forward_gelu_f32(
const struct ggml_tensor * src0 = dst->src[0]; const struct ggml_tensor * src0 = dst->src[0];
GGML_ASSERT(ggml_is_contiguous_except_dim_1(src0)); GGML_ASSERT(ggml_is_contiguous_1(src0));
GGML_ASSERT(ggml_is_contiguous_except_dim_1(dst)); GGML_ASSERT(ggml_is_contiguous_1(dst));
GGML_ASSERT(ggml_are_same_shape(src0, dst)); GGML_ASSERT(ggml_are_same_shape(src0, dst));
if (params->type == GGML_TASK_TYPE_INIT || params->type == GGML_TASK_TYPE_FINALIZE) { if (params->type == GGML_TASK_TYPE_INIT || params->type == GGML_TASK_TYPE_FINALIZE) {
@ -11483,8 +11495,8 @@ static void ggml_compute_forward_gelu_quick_f32(
const struct ggml_tensor * src0 = dst->src[0]; const struct ggml_tensor * src0 = dst->src[0];
GGML_ASSERT(ggml_is_contiguous_except_dim_1(src0)); GGML_ASSERT(ggml_is_contiguous_1(src0));
GGML_ASSERT(ggml_is_contiguous_except_dim_1(dst)); GGML_ASSERT(ggml_is_contiguous_1(dst));
GGML_ASSERT(ggml_are_same_shape(src0, dst)); GGML_ASSERT(ggml_are_same_shape(src0, dst));
if (params->type == GGML_TASK_TYPE_INIT || params->type == GGML_TASK_TYPE_FINALIZE) { if (params->type == GGML_TASK_TYPE_INIT || params->type == GGML_TASK_TYPE_FINALIZE) {
@ -11546,8 +11558,8 @@ static void ggml_compute_forward_silu_f32(
const struct ggml_tensor * src0 = dst->src[0]; const struct ggml_tensor * src0 = dst->src[0];
GGML_ASSERT(ggml_is_contiguous_except_dim_1(src0)); GGML_ASSERT(ggml_is_contiguous_1(src0));
GGML_ASSERT(ggml_is_contiguous_except_dim_1(dst)); GGML_ASSERT(ggml_is_contiguous_1(dst));
GGML_ASSERT(ggml_are_same_shape(src0, dst)); GGML_ASSERT(ggml_are_same_shape(src0, dst));
if (params->type == GGML_TASK_TYPE_INIT || params->type == GGML_TASK_TYPE_FINALIZE) { if (params->type == GGML_TASK_TYPE_INIT || params->type == GGML_TASK_TYPE_FINALIZE) {
@ -11658,9 +11670,9 @@ static void ggml_compute_forward_silu_back_f32(
const struct ggml_tensor * src0 = dst->src[0]; const struct ggml_tensor * src0 = dst->src[0];
const struct ggml_tensor * grad = dst->src[1]; const struct ggml_tensor * grad = dst->src[1];
GGML_ASSERT(ggml_is_contiguous_except_dim_1(grad)); GGML_ASSERT(ggml_is_contiguous_1(grad));
GGML_ASSERT(ggml_is_contiguous_except_dim_1(src0)); GGML_ASSERT(ggml_is_contiguous_1(src0));
GGML_ASSERT(ggml_is_contiguous_except_dim_1(dst)); GGML_ASSERT(ggml_is_contiguous_1(dst));
GGML_ASSERT(ggml_are_same_shape(src0, dst)); GGML_ASSERT(ggml_are_same_shape(src0, dst));
GGML_ASSERT(ggml_are_same_shape(src0, grad)); GGML_ASSERT(ggml_are_same_shape(src0, grad));
@ -14358,7 +14370,7 @@ static void ggml_compute_forward_rope_f32(
int ir = 0; int ir = 0;
const float theta_scale = powf(freq_base, -2.0f/n_dims); const float theta_scale = powf(freq_base, -2.0f/n_dims);
const float inv_ndims = -1.f/n_dims;
float corr_dims[2]; float corr_dims[2];
ggml_rope_yarn_corr_dims(n_dims, n_orig_ctx, freq_base, beta_fast, beta_slow, corr_dims); ggml_rope_yarn_corr_dims(n_dims, n_orig_ctx, freq_base, beta_fast, beta_slow, corr_dims);
@ -14442,29 +14454,22 @@ static void ggml_compute_forward_rope_f32(
dst_data[1] = x0*sin_theta*zeta + x1*cos_theta*zeta; dst_data[1] = x0*sin_theta*zeta + x1*cos_theta*zeta;
} }
} else { } else {
// TODO: this might be wrong for ne0 != n_dims - need double check // ref: https://github.com/jquesnelle/yarn/blob/master/scaled_rope/LlamaYaRNScaledRotaryEmbedding.py
// it seems we have to rope just the first n_dims elements and do nothing with the rest
// ref: https://github.com/ml-explore/mlx/blob/dc2edc762c797e3b8de50b1dad4dc0a131691033/benchmarks/python/llama_jax_bench.py#L11-L26
theta_base *= freq_scale;
for (int64_t ic = 0; ic < ne0; ic += 2) { for (int64_t ic = 0; ic < ne0; ic += 2) {
if (ic < n_dims) { if (ic < n_dims) {
const int64_t ib = 0; const int64_t i0 = ic/2;
// simplified from `(ib * n_dims + ic) * inv_ndims` const float freq_factor = freq_factors ? freq_factors[i0] : 1.0f;
float cur_rot = inv_ndims * ic - ib;
float freq_factor = freq_factors ? freq_factors[ic/2] : 1.0f;
float cos_theta, sin_theta; float cos_theta, sin_theta;
rope_yarn( rope_yarn(
theta_base/freq_factor, freq_scale, corr_dims, cur_rot, ext_factor, attn_factor, theta_base/freq_factor, freq_scale, corr_dims, ic, ext_factor, attn_factor,
&cos_theta, &sin_theta &cos_theta, &sin_theta
); );
sin_theta *= sin_sign; sin_theta *= sin_sign;
theta_base *= theta_scale; theta_base *= theta_scale;
const int64_t i0 = ib*n_dims + ic/2;
const float * const src = (float *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); const float * const src = (float *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
float * dst_data = (float *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); float * dst_data = (float *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
@ -14543,7 +14548,7 @@ static void ggml_compute_forward_rope_f16(
int ir = 0; int ir = 0;
const float theta_scale = powf(freq_base, -2.0f/n_dims); const float theta_scale = powf(freq_base, -2.0f/n_dims);
const float inv_ndims = -1.f/n_dims;
float corr_dims[2]; float corr_dims[2];
ggml_rope_yarn_corr_dims(n_dims, n_orig_ctx, freq_base, beta_fast, beta_slow, corr_dims); ggml_rope_yarn_corr_dims(n_dims, n_orig_ctx, freq_base, beta_fast, beta_slow, corr_dims);
@ -14623,29 +14628,22 @@ static void ggml_compute_forward_rope_f16(
dst_data[1] = GGML_FP32_TO_FP16(x0*sin_theta + x1*cos_theta); dst_data[1] = GGML_FP32_TO_FP16(x0*sin_theta + x1*cos_theta);
} }
} else { } else {
// TODO: this might be wrong for ne0 != n_dims - need double check // ref: https://github.com/jquesnelle/yarn/blob/master/scaled_rope/LlamaYaRNScaledRotaryEmbedding.py
// it seems we have to rope just the first n_dims elements and do nothing with the rest
// ref: https://github.com/ml-explore/mlx/blob/dc2edc762c797e3b8de50b1dad4dc0a131691033/benchmarks/python/llama_jax_bench.py#L11-L26
theta_base *= freq_scale;
for (int64_t ic = 0; ic < ne0; ic += 2) { for (int64_t ic = 0; ic < ne0; ic += 2) {
if (ic < n_dims) { if (ic < n_dims) {
const int64_t ib = 0; const int64_t i0 = ic/2;
// simplified from `(ib * n_dims + ic) * inv_ndims` const float freq_factor = freq_factors ? freq_factors[i0] : 1.0f;
float cur_rot = inv_ndims * ic - ib;
float freq_factor = freq_factors ? freq_factors[ic/2] : 1.0f;
float cos_theta, sin_theta; float cos_theta, sin_theta;
rope_yarn( rope_yarn(
theta_base/freq_factor, freq_scale, corr_dims, cur_rot, ext_factor, attn_factor, theta_base/freq_factor, freq_scale, corr_dims, ic, ext_factor, attn_factor,
&cos_theta, &sin_theta &cos_theta, &sin_theta
); );
sin_theta *= sin_sign; sin_theta *= sin_sign;
theta_base *= theta_scale; theta_base *= theta_scale;
const int64_t i0 = ib*n_dims + ic/2;
const ggml_fp16_t * const src = (ggml_fp16_t *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); const ggml_fp16_t * const src = (ggml_fp16_t *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
ggml_fp16_t * dst_data = (ggml_fp16_t *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); ggml_fp16_t * dst_data = (ggml_fp16_t *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);

6
ggml.h
View File

@ -756,7 +756,6 @@ extern "C" {
GGML_API enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype); GGML_API enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype);
GGML_API GGML_CALL bool ggml_is_transposed(const struct ggml_tensor * tensor); GGML_API GGML_CALL bool ggml_is_transposed(const struct ggml_tensor * tensor);
GGML_API GGML_CALL bool ggml_is_contiguous(const struct ggml_tensor * tensor);
GGML_API GGML_CALL bool ggml_is_permuted (const struct ggml_tensor * tensor); GGML_API GGML_CALL bool ggml_is_permuted (const struct ggml_tensor * tensor);
GGML_API GGML_CALL bool ggml_is_empty (const struct ggml_tensor * tensor); GGML_API GGML_CALL bool ggml_is_empty (const struct ggml_tensor * tensor);
GGML_API bool ggml_is_scalar (const struct ggml_tensor * tensor); GGML_API bool ggml_is_scalar (const struct ggml_tensor * tensor);
@ -765,6 +764,11 @@ extern "C" {
GGML_API bool ggml_is_3d (const struct ggml_tensor * tensor); GGML_API bool ggml_is_3d (const struct ggml_tensor * tensor);
GGML_API int ggml_n_dims (const struct ggml_tensor * tensor); // returns 1 for scalars GGML_API int ggml_n_dims (const struct ggml_tensor * tensor); // returns 1 for scalars
GGML_API GGML_CALL bool ggml_is_contiguous (const struct ggml_tensor * tensor);
GGML_API GGML_CALL bool ggml_is_contiguous_0(const struct ggml_tensor * tensor); // same as ggml_is_contiguous()
GGML_API GGML_CALL bool ggml_is_contiguous_1(const struct ggml_tensor * tensor); // contiguous for dims >= 1
GGML_API GGML_CALL bool ggml_is_contiguous_2(const struct ggml_tensor * tensor); // contiguous for dims >= 2
GGML_API bool ggml_are_same_shape (const struct ggml_tensor * t0, const struct ggml_tensor * t1); GGML_API bool ggml_are_same_shape (const struct ggml_tensor * t0, const struct ggml_tensor * t1);
GGML_API bool ggml_are_same_stride(const struct ggml_tensor * t0, const struct ggml_tensor * t1); GGML_API bool ggml_are_same_stride(const struct ggml_tensor * t0, const struct ggml_tensor * t1);

4
ggml_vk_generate_shaders.py
View File

@ -2670,14 +2670,12 @@ void main() {
const uint i = row*p.ncols + ib*p.ndims + ic/2; const uint i = row*p.ncols + ib*p.ndims + ic/2;
const uint i2 = row/p.p_delta_rows; const uint i2 = row/p.p_delta_rows;
const float cur_rot = p.inv_ndims * ic - ib;
const int pos = data_b[i2]; const int pos = data_b[i2];
const float freq_factor = p.has_freq_facs != 0 ? data_freq_factors[ic/2] : 1.0f; const float freq_factor = p.has_freq_facs != 0 ? data_freq_factors[ic/2] : 1.0f;
const float theta_base = pos*p.freq_scale*pow(p.theta_scale, col/2.0f) / freq_factor; const float theta_base = pos*p.freq_scale*pow(p.theta_scale, col/2.0f) / freq_factor;
float cos_theta, sin_theta; float cos_theta, sin_theta;
rope_yarn(theta_base, uint(cur_rot), cos_theta, sin_theta); rope_yarn(theta_base, ic, cos_theta, sin_theta);
const float x0 = float(data_a[i + 0]); const float x0 = float(data_a[i + 0]);
const float x1 = float(data_a[i + p.ndims/2]); const float x1 = float(data_a[i + p.ndims/2]);

52
llama.cpp
View File

@ -11187,46 +11187,69 @@ struct llm_build_context {
} }
// split into {n_head * n_embd_head_qk_nope, n_tokens} // split into {n_head * n_embd_head_qk_nope, n_tokens}
struct ggml_tensor * q_nope = ggml_view_3d(ctx0, q, n_embd_head_qk_nope, n_head, n_tokens, ggml_element_size(q) * hparams.n_embd_head_k, ggml_element_size(q) * hparams.n_embd_head_k * n_head, 0); struct ggml_tensor * q_nope = ggml_view_3d(ctx0, q, n_embd_head_qk_nope, n_head, n_tokens,
ggml_row_size(q->type, hparams.n_embd_head_k),
ggml_row_size(q->type, hparams.n_embd_head_k * n_head),
0);
cb(q_nope, "q_nope", il); cb(q_nope, "q_nope", il);
// and {n_head * n_embd_head_qk_rope, n_tokens} // and {n_head * n_embd_head_qk_rope, n_tokens}
struct ggml_tensor * q_pe = ggml_view_3d(ctx0, q, n_embd_head_qk_rope, n_head, n_tokens, ggml_element_size(q) * hparams.n_embd_head_k, ggml_element_size(q) * hparams.n_embd_head_k * n_head, ggml_element_size(q) * n_embd_head_qk_nope); struct ggml_tensor * q_pe = ggml_view_3d(ctx0, q, n_embd_head_qk_rope, n_head, n_tokens,
ggml_row_size(q->type, hparams.n_embd_head_k),
ggml_row_size(q->type, hparams.n_embd_head_k * n_head),
ggml_row_size(q->type, n_embd_head_qk_nope));
cb(q_pe, "q_pe", il); cb(q_pe, "q_pe", il);
// {n_embd, kv_lora_rank + n_embd_head_qk_rope} * {n_embd, n_tokens} -> {kv_lora_rank + n_embd_head_qk_rope, n_tokens} // {n_embd, kv_lora_rank + n_embd_head_qk_rope} * {n_embd, n_tokens} -> {kv_lora_rank + n_embd_head_qk_rope, n_tokens}
struct ggml_tensor * compressed_kv_pe = ggml_mul_mat(ctx0, model.layers[il].wkv_a_mqa, cur); struct ggml_tensor * kv_pe_compresseed = ggml_mul_mat(ctx0, model.layers[il].wkv_a_mqa, cur);
cb(compressed_kv_pe, "compressed_kv_pe", il); cb(kv_pe_compresseed, "kv_pe_compresseed", il);
// split into {kv_lora_rank, n_tokens} // split into {kv_lora_rank, n_tokens}
struct ggml_tensor * compressed_kv = ggml_view_2d(ctx0, compressed_kv_pe, kv_lora_rank, n_tokens, compressed_kv_pe->nb[1], 0); struct ggml_tensor * kv_compressed = ggml_view_2d(ctx0, kv_pe_compresseed, kv_lora_rank, n_tokens,
cb(compressed_kv, "compressed_kv", il); kv_pe_compresseed->nb[1],
0);
cb(kv_compressed, "kv_compressed", il);
// and {n_embd_head_qk_rope, n_tokens} // and {n_embd_head_qk_rope, n_tokens}
struct ggml_tensor * k_pe = ggml_view_2d(ctx0, compressed_kv_pe, n_embd_head_qk_rope, n_tokens, compressed_kv_pe->nb[1], ggml_element_size(compressed_kv_pe)*kv_lora_rank); struct ggml_tensor * k_pe = ggml_view_3d(ctx0, kv_pe_compresseed, n_embd_head_qk_rope, 1, n_tokens,
kv_pe_compresseed->nb[1],
kv_pe_compresseed->nb[1],
ggml_row_size(kv_pe_compresseed->type, kv_lora_rank));
cb(k_pe, "k_pe", il); cb(k_pe, "k_pe", il);
compressed_kv = llm_build_norm(ctx0, compressed_kv, hparams, kv_compressed = ggml_cont(ctx0, kv_compressed); // TODO: the CUDA backend does not support non-contiguous norm
kv_compressed = llm_build_norm(ctx0, kv_compressed, hparams,
model.layers[il].attn_kv_a_norm, NULL, model.layers[il].attn_kv_a_norm, NULL,
LLM_NORM_RMS, cb, il); LLM_NORM_RMS, cb, il);
cb(compressed_kv, "compressed_kv", il); cb(kv_compressed, "kv_compressed", il);
// {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v)} * {kv_lora_rank, n_tokens} -> {n_head * (n_embd_head_qk_nope + n_embd_head_v), n_tokens} // {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v)} * {kv_lora_rank, n_tokens} -> {n_head * (n_embd_head_qk_nope + n_embd_head_v), n_tokens}
struct ggml_tensor * kv = ggml_mul_mat(ctx0, model.layers[il].wkv_b, compressed_kv); struct ggml_tensor * kv = ggml_mul_mat(ctx0, model.layers[il].wkv_b, kv_compressed);
cb(kv, "kv", il); cb(kv, "kv", il);
// split into {n_head * n_embd_head_qk_nope, n_tokens} // split into {n_head * n_embd_head_qk_nope, n_tokens}
struct ggml_tensor * k_nope = ggml_view_3d(ctx0, kv, n_embd_head_qk_nope, n_head, n_tokens, ggml_element_size(kv) * (n_embd_head_qk_nope + hparams.n_embd_head_v), ggml_element_size(kv) * n_head * (n_embd_head_qk_nope + hparams.n_embd_head_v), 0); struct ggml_tensor * k_nope = ggml_view_3d(ctx0, kv, n_embd_head_qk_nope, n_head, n_tokens,
ggml_row_size(kv->type, n_embd_head_qk_nope + hparams.n_embd_head_v),
ggml_row_size(kv->type, n_head * (n_embd_head_qk_nope + hparams.n_embd_head_v)),
0);
cb(k_nope, "k_nope", il); cb(k_nope, "k_nope", il);
// and {n_head * n_embd_head_v, n_tokens} // and {n_head * n_embd_head_v, n_tokens}
struct ggml_tensor * v_states = ggml_view_3d(ctx0, kv, hparams.n_embd_head_v, n_head, n_tokens, ggml_element_size(kv) * (n_embd_head_qk_nope + hparams.n_embd_head_v), ggml_element_size(kv) * n_head * (n_embd_head_qk_nope + hparams.n_embd_head_v), ggml_element_size(kv) * n_embd_head_qk_nope); struct ggml_tensor * v_states = ggml_view_3d(ctx0, kv, hparams.n_embd_head_v, n_head, n_tokens,
ggml_row_size(kv->type, (n_embd_head_qk_nope + hparams.n_embd_head_v)),
ggml_row_size(kv->type, (n_embd_head_qk_nope + hparams.n_embd_head_v)*n_head),
ggml_row_size(kv->type, (n_embd_head_qk_nope)));
cb(v_states, "v_states", il); cb(v_states, "v_states", il);
v_states = ggml_cont(ctx0, v_states); v_states = ggml_cont(ctx0, v_states);
cb(v_states, "v_states", il); cb(v_states, "v_states", il);
v_states = ggml_view_2d(ctx0, v_states, hparams.n_embd_head_v * n_head, n_tokens, ggml_element_size(kv) * hparams.n_embd_head_v * n_head, 0); v_states = ggml_view_2d(ctx0, v_states, hparams.n_embd_head_v * n_head, n_tokens,
ggml_row_size(kv->type, hparams.n_embd_head_v * n_head),
0);
cb(v_states, "v_states", il); cb(v_states, "v_states", il);
q_pe = ggml_cont(ctx0, q_pe); // TODO: the CUDA backend does not support non-contiguous RoPE
q_pe = ggml_rope_ext( q_pe = ggml_rope_ext(
ctx0, q_pe, inp_pos, nullptr, ctx0, q_pe, inp_pos, nullptr,
n_rot, rope_type, 0, n_orig_ctx, freq_base, freq_scale, n_rot, rope_type, 0, n_orig_ctx, freq_base, freq_scale,
@ -11235,8 +11258,9 @@ struct llm_build_context {
cb(q_pe, "q_pe", il); cb(q_pe, "q_pe", il);
// shared RoPE key // shared RoPE key
k_pe = ggml_cont(ctx0, k_pe); // TODO: the CUDA backend does not support non-contiguous RoPE
k_pe = ggml_rope_ext( k_pe = ggml_rope_ext(
ctx0, ggml_view_3d(ctx0, k_pe, n_embd_head_qk_rope, 1, n_tokens, k_pe->nb[0], k_pe->nb[1], 0), inp_pos, nullptr, ctx0, k_pe, inp_pos, nullptr,
n_rot, rope_type, 0, n_orig_ctx, freq_base, freq_scale, n_rot, rope_type, 0, n_orig_ctx, freq_base, freq_scale,
ext_factor, attn_factor_scaled, beta_fast, beta_slow ext_factor, attn_factor_scaled, beta_fast, beta_slow
); );

71
tests/test-backend-ops.cpp
View File

@ -1138,26 +1138,37 @@ struct test_soft_max : public test_case {
// GGML_OP_ROPE // GGML_OP_ROPE
struct test_rope : public test_case { struct test_rope : public test_case {
const ggml_type type; const ggml_type type;
const std::array<int64_t, 4> ne; const std::array<int64_t, 4> ne_a;
int n_dims; int n_dims;
int mode; int mode;
int n_ctx; int n_ctx;
float fs; // freq_scale
float ef; // ext_factor
float af; // attn_factor
bool ff; bool ff;
int v; // view (1 : non-contiguous a)
std::string vars() override { std::string vars() override {
return VARS_TO_STR6(type, ne, n_dims, mode, n_ctx, ff); return VARS_TO_STR10(type, ne_a, n_dims, mode, n_ctx, fs, ef, af, ff, v);
} }
test_rope(ggml_type type = GGML_TYPE_F32, test_rope(ggml_type type = GGML_TYPE_F32,
std::array<int64_t, 4> ne = {10, 10, 10, 1}, std::array<int64_t, 4> ne_a = {10, 10, 10, 1},
int n_dims = 10, int mode = 0, int n_ctx = 512, bool ff = false) int n_dims = 10, int mode = 0, int n_ctx = 512, float fs = 1.0f, float ef = 0.0f, float af = 0.0f, bool ff = false, int v = 0)
: type(type), ne(ne), n_dims(n_dims), mode(mode), n_ctx(n_ctx), ff(ff) {} : type(type), ne_a(ne_a), n_dims(n_dims), mode(mode), n_ctx(n_ctx), fs(fs), ef(ef), af(af), ff(ff), v(v) {}
ggml_tensor * build_graph(ggml_context * ctx) override { ggml_tensor * build_graph(ggml_context * ctx) override {
ggml_tensor * a = ggml_new_tensor(ctx, type, 4, ne.data()); ggml_tensor * a;
ggml_tensor * pos = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, ne[2]); if (v & 1) {
auto ne = ne_a; ne[0] *= 2; ne[1] *= 4; ne[2] *= 3;
a = ggml_new_tensor(ctx, type, 4, ne.data());
a = ggml_view_4d(ctx, a, ne_a[0], ne_a[1], ne_a[2], ne_a[3], a->nb[1], a->nb[2], a->nb[3], 0);
} else {
a = ggml_new_tensor(ctx, type, 4, ne_a.data());
}
ggml_tensor * pos = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, ne_a[2]);
ggml_tensor * freq = ff ? ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_dims/2) : nullptr; ggml_tensor * freq = ff ? ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_dims/2) : nullptr;
ggml_tensor * out = ggml_rope_ext(ctx, a, pos, freq, n_dims, mode, n_ctx, 0, 10000.0f, 1.0f, 0.0f, 1.0f, 0.0f, 0.0f); ggml_tensor * out = ggml_rope_ext(ctx, a, pos, freq, n_dims, mode, n_ctx, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
return out; return out;
} }
@ -1165,11 +1176,11 @@ struct test_rope : public test_case {
for (ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) { for (ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) {
if (t->type == GGML_TYPE_I32) { if (t->type == GGML_TYPE_I32) {
// pos // pos
std::vector<int> data(ne[2]); std::vector<int> data(ne_a[2]);
for (int i = 0; i < ne[2]; i++) { for (int i = 0; i < ne_a[2]; i++) {
data[i] = rand() % n_ctx; data[i] = rand() % n_ctx;
} }
ggml_backend_tensor_set(t, data.data(), 0, ne[2] * sizeof(int)); ggml_backend_tensor_set(t, data.data(), 0, ne_a[2] * sizeof(int));
} else { } else {
if (t->ne[0] == n_dims/2) { if (t->ne[0] == n_dims/2) {
// frequency factors in the range [0.9f, 1.1f] // frequency factors in the range [0.9f, 1.1f]
@ -2213,20 +2224,38 @@ static bool test_backend(ggml_backend_t backend, test_mode mode, const char * op
test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {32, 2, 32, 1}, true, 0.1f, 0.0f)); test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {32, 2, 32, 1}, true, 0.1f, 0.0f));
test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {32, 2, 32, 1}, true, 0.1f, 8.0f)); test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {32, 2, 32, 1}, true, 0.1f, 8.0f));
{
bool all = true;
for (float v : { 0, 1 }) {
for (float fs : { 1.0f, 1.4245f }) {
for (float ef : { 0.0f, 0.7465f }) {
for (float af : { 1.0f, 1.4245f }) {
for (ggml_type type : {GGML_TYPE_F32, GGML_TYPE_F16}) { for (ggml_type type : {GGML_TYPE_F32, GGML_TYPE_F16}) {
// TODO: ff not supported yet for !neox // TODO: ff not supported yet for !neox
test_cases.emplace_back(new test_rope(type, {128, 32, 10, 1}, 128, 0, 512, false)); // llama 7B test_cases.emplace_back(new test_rope(type, {128, 32, 10, 1}, 128, 0, 512, fs, ef, af, false, v)); // llama 7B
test_cases.emplace_back(new test_rope(type, {128, 40, 10, 1}, 128, 0, 512, false)); // llama 13B if (all) {
test_cases.emplace_back(new test_rope(type, {128, 52, 10, 1}, 128, 0, 512, false)); // llama 30B test_cases.emplace_back(new test_rope(type, {128, 40, 10, 1}, 128, 0, 512, fs, ef, af, false, v)); // llama 13B
test_cases.emplace_back(new test_rope(type, {128, 64, 10, 1}, 128, 0, 512, false)); // llama 65B test_cases.emplace_back(new test_rope(type, {128, 52, 10, 1}, 128, 0, 512, fs, ef, af, false, v)); // llama 30B
test_cases.emplace_back(new test_rope(type, {128, 64, 10, 1}, 128, 0, 512, fs, ef, af, false, v)); // llama 65B
}
for (bool ff : {false, true}) { // freq_factors for (bool ff : {false, true}) { // freq_factors
test_cases.emplace_back(new test_rope(type, { 64, 1, 10, 1}, 64, 2, 512, ff)); // neox (falcon 7B) if (all) {
test_cases.emplace_back(new test_rope(type, { 64, 71, 10, 1}, 64, 2, 512, ff)); // neox (falcon 7B) test_cases.emplace_back(new test_rope(type, { 64, 1, 10, 1}, 64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 7B)
test_cases.emplace_back(new test_rope(type, { 64, 8, 10, 1}, 64, 2, 512, ff)); // neox (falcon 40B) test_cases.emplace_back(new test_rope(type, { 64, 71, 10, 1}, 64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 7B)
test_cases.emplace_back(new test_rope(type, { 64, 128, 10, 1}, 64, 2, 512, ff)); // neox (falcon 40B) test_cases.emplace_back(new test_rope(type, { 64, 8, 10, 1}, 64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 40B)
test_cases.emplace_back(new test_rope(type, { 80, 32, 10, 1}, 20, 2, 512, ff)); // neox (stablelm) test_cases.emplace_back(new test_rope(type, { 80, 32, 10, 1}, 20, 2, 512, fs, ef, af, ff, v)); // neox (stablelm)
test_cases.emplace_back(new test_rope(type, { 80, 32, 10, 1}, 32, 2, 512, ff)); // neox (phi-2) test_cases.emplace_back(new test_rope(type, { 80, 32, 10, 1}, 32, 2, 512, fs, ef, af, ff, v)); // neox (phi-2)
}
test_cases.emplace_back(new test_rope(type, { 64, 128, 10, 1}, 64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 40B)
}
}
all = false;
}
}
}
} }
} }