Depthwise 2D convolution

include/ggml.h

@@ -481,6 +481,7 @@ extern "C" {
         GGML_OP_CONV_TRANSPOSE_1D,
         GGML_OP_IM2COL,
         GGML_OP_IM2COL_BACK,
+        GGML_OP_DEPTHWISE_CONV_2D,
         GGML_OP_CONV_TRANSPOSE_2D,
         GGML_OP_POOL_1D,
         GGML_OP_POOL_2D,
@@ -677,6 +678,9 @@ extern "C" {
     GGML_API bool ggml_is_contiguous_1(const struct ggml_tensor * tensor); // contiguous for dims >= 1
     GGML_API bool ggml_is_contiguous_2(const struct ggml_tensor * tensor); // contiguous for dims >= 2
 
+    // true for tensor that is stored in memory as CxWxHxN and has been permuted to WxHxCxN
+    GGML_API bool ggml_is_contiguous_channels(const struct ggml_tensor * tensor);
+
     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);
 
@@ -1660,7 +1664,7 @@ extern "C" {
             struct ggml_tensor  * a,
             struct ggml_tensor  * b);
 
-    // depthwise
+    // depthwise (via im2col and mul_mat)
     GGML_API struct ggml_tensor * ggml_conv_2d_dw(
             struct ggml_context * ctx,
             struct ggml_tensor  * a,  // convolution kernel
@@ -1672,6 +1676,19 @@ extern "C" {
             int                  d0,  // dilation dimension 0
             int                  d1); // dilation dimension 1
 
+    // Depthwise 2D convolution
+    // a:   KW    KH    1    C    convolution kernel
+    // b:   W     H     C    N    input data
+    // res: W_out H_out C    N
+    GGML_API struct ggml_tensor * ggml_depthwise_conv_2d(
+            struct ggml_context * ctx,
+            struct ggml_tensor  * a,
+            struct ggml_tensor  * b,
+            int                   stride0,
+            int                   stride1,
+            int                   pad0,
+            int                   pad1);
+
     GGML_API struct ggml_tensor * ggml_conv_transpose_2d_p0(
             struct ggml_context * ctx,
             struct ggml_tensor  * a,

src/ggml-cpu/ggml-cpu.c

@@ -1932,6 +1932,10 @@ static void ggml_compute_forward(struct ggml_compute_params * params, struct ggm
             {
                 ggml_compute_forward_im2col_back_f32(params, tensor);
             } break;
+        case GGML_OP_DEPTHWISE_CONV_2D:
+            {
+                ggml_compute_forward_depthwise_conv_2d(params, tensor);
+            } break;
         case GGML_OP_CONV_TRANSPOSE_2D:
             {
                 ggml_compute_forward_conv_transpose_2d(params, tensor);
@@ -2268,6 +2272,7 @@ static int ggml_get_n_tasks(struct ggml_tensor * node, int n_threads) {
             } break;
         case GGML_OP_IM2COL:
         case GGML_OP_IM2COL_BACK:
+        case GGML_OP_DEPTHWISE_CONV_2D:
         case GGML_OP_CONV_TRANSPOSE_1D:
         case GGML_OP_CONV_TRANSPOSE_2D:
             {

src/ggml-cpu/ops.cpp

@@ -6064,6 +6064,174 @@ void ggml_compute_forward_conv_transpose_2d(
     }
 }
 
+// ggml_compute_forward_depthwise_conv_2d
+
+struct ggml_depthwise_conv_2d_params {
+    int64_t channels;
+    int64_t batch;
+    int64_t src_w;
+    int64_t src_h;
+    int64_t dst_w;
+    int64_t dst_h;
+    int64_t knl_w;
+    int64_t knl_h;
+    int stride_x;
+    int stride_y;
+    int pad_x;
+    int pad_y;
+};
+
+static void ggml_compute_forward_depthwise_conv_2d_cwhn(
+        const struct ggml_compute_params * params,
+        const struct ggml_tensor * src,
+        const struct ggml_tensor * kernel,
+        struct ggml_tensor * dst,
+        const struct ggml_depthwise_conv_2d_params p) {
+
+    const int64_t c = p.channels;
+    const float * knl_data = (const float *)kernel->data;
+
+    const int64_t rows_total = p.dst_h * p.batch;
+    const int64_t rows_per_thread = (rows_total + params->nth - 1) / params->nth;
+    const int64_t row_start = params->ith * rows_per_thread;
+    const int64_t row_end = MIN(row_start + rows_per_thread, rows_total);
+
+#ifdef GGML_SIMD
+    const int64_t pkg_size = GGML_F32_EPR;
+    const int64_t pkg_count = c / pkg_size;
+    const int64_t c_pkg_end = pkg_count * pkg_size;
+#else
+    const int64_t c_pkg_end = 0;
+#endif
+
+    for (int64_t row = row_start; row < row_end; ++row) {
+        const int64_t dst_y = row % p.dst_h;
+        const float * src_data = (const float *)src->data + (row / p.dst_h) * p.src_w * p.src_h * c;
+        for (int64_t dst_x = 0; dst_x < p.dst_w; ++dst_x) {
+            float * dst_data = (float *)dst->data + (row * p.dst_w + dst_x) * c;
+            const int64_t src_y_base = dst_y * p.stride_y - p.pad_y;
+            const int64_t src_x_base = dst_x * p.stride_x - p.pad_x;
+
+#ifdef GGML_SIMD
+            // Vectorized loop
+            for (int64_t c_i = 0; c_i < c_pkg_end; c_i += pkg_size) {
+                GGML_F32_VEC sum = GGML_F32_VEC_ZERO;
+                for (int64_t knl_y = 0; knl_y < p.knl_h; ++knl_y) {
+                    const int64_t src_y = src_y_base + knl_y;
+                    if (src_y < 0 || src_y >= p.src_h) {
+                        continue;
+                    }
+                    for (int64_t knl_x = 0; knl_x < p.knl_w; ++knl_x) {
+                        const int64_t src_x = src_x_base + knl_x;
+                        if (src_x < 0 || src_x >= p.src_w) {
+                            continue;
+                        }
+                        GGML_F32_VEC k = GGML_F32_VEC_LOAD(knl_data + (knl_y * p.knl_w + knl_x) * c + c_i);
+                        GGML_F32_VEC s = GGML_F32_VEC_LOAD(src_data + (src_y * p.src_w + src_x) * c + c_i);
+                        sum = GGML_F32_VEC_FMA(sum, k, s);
+                    }
+                }
+                GGML_F32_VEC_STORE(dst_data + c_i, sum);
+            }
+#endif
+            // Scalar loop
+            for (int64_t c_i = c_pkg_end; c_i < c; ++c_i) {
+                float sum = 0.0f;
+                for (int64_t knl_y = 0; knl_y < p.knl_h; ++knl_y) {
+                    const int64_t src_y = src_y_base + knl_y;
+                    if (src_y < 0 || src_y >= p.src_h) {
+                        continue;
+                    }
+                    for (int64_t knl_x = 0; knl_x < p.knl_w; ++knl_x) {
+                        const int64_t src_x = src_x_base + knl_x;
+                        if (src_x < 0 || src_x >= p.src_w) {
+                            continue;
+                        }
+                        sum += knl_data[(knl_y * p.knl_w + knl_x) * c + c_i]
+                             * src_data[(src_y * p.src_w + src_x) * c + c_i];
+                    }
+                }
+                dst_data[c_i] = sum;
+            }
+        }
+    }
+}
+
+static void ggml_compute_forward_depthwise_conv_2d_whcn(
+        const struct ggml_compute_params * params,
+        const struct ggml_tensor * src,
+        const struct ggml_tensor * kernel,
+        struct ggml_tensor * dst,
+        const struct ggml_depthwise_conv_2d_params p) {
+
+    const int64_t n = p.channels * p.batch;
+    const int64_t per_thread = (n + params->nth - 1) / params->nth;
+    const int64_t start = params->ith * per_thread;
+    const int64_t end = MIN(start + per_thread, n);
+
+    for (int64_t i = start; i < end; ++i) {
+        const float * knl_data = (const float *)kernel->data + (i % p.channels) * p.knl_w * p.knl_h;
+        const float * src_data = (const float *)src->data + i * p.src_w * p.src_h;
+        float * dst_data = (float *)dst->data + i * p.dst_w * p.dst_h;
+
+        for (int64_t dst_y = 0; dst_y < p.dst_h; ++dst_y) {
+            for (int64_t dst_x = 0; dst_x < p.dst_w; ++dst_x) {
+
+                float sum = 0.0f;                
+                for (int64_t knl_y = 0; knl_y < p.knl_h; ++knl_y) {
+                    const int64_t src_y = dst_y * p.stride_y + knl_y - p.pad_y;
+                    if (src_y < 0 || src_y >= p.src_h) {
+                        continue;
+                    }
+                    for (int64_t knl_x = 0; knl_x < p.knl_w; ++knl_x) {
+                        const int64_t src_x = dst_x * p.stride_x + knl_x - p.pad_x;
+                        if (src_x < 0 || src_x >= p.src_w) {
+                            continue;
+                        }
+                        sum += knl_data[knl_y * p.knl_w + knl_x]
+                             * src_data[src_y * p.src_w + src_x];
+                    }
+                }
+                dst_data[dst_y * p.dst_w + dst_x] = sum;
+            }
+        }
+    }
+}
+
+void ggml_compute_forward_depthwise_conv_2d(
+        const struct ggml_compute_params * params,
+        struct ggml_tensor * dst) {
+
+    const struct ggml_tensor * kernel = dst->src[0];
+    const struct ggml_tensor * src = dst->src[1];
+    struct ggml_depthwise_conv_2d_params p;
+    p.channels = src->ne[2];
+    p.batch = src->ne[3];
+    p.src_w = src->ne[0];
+    p.src_h = src->ne[1];
+    p.dst_w = dst->ne[0];
+    p.dst_h = dst->ne[1];
+    p.knl_w = kernel->ne[0];
+    p.knl_h = kernel->ne[1];
+    p.stride_x = dst->op_params[0];
+    p.stride_y = dst->op_params[1];
+    p.pad_x = dst->op_params[2];
+    p.pad_y = dst->op_params[3];
+
+    GGML_ASSERT(kernel->ne[3] == p.channels);
+    GGML_ASSERT(dst->ne[3] == p.batch);
+
+    if (ggml_is_contiguous(src)) {
+        ggml_compute_forward_depthwise_conv_2d_whcn(params, src, kernel, dst, p);
+    } else if (ggml_is_contiguous_channels(src)) {
+        // kernel should also have channels most contiguous in memory
+        GGML_ASSERT(kernel->nb[0] >= kernel->nb[2] && kernel->nb[1] >= kernel->nb[0]);
+        ggml_compute_forward_depthwise_conv_2d_cwhn(params, src, kernel, dst, p);
+    } else {
+        GGML_ABORT("non-contiguous memory layout not supported");
+    }
+}
+
 // ggml_compute_forward_pool_1d_sk_p0
 
 static void ggml_compute_forward_pool_1d_sk_p0(

src/ggml-cpu/ops.h

@@ -65,6 +65,7 @@ void ggml_compute_forward_conv_transpose_1d(const struct ggml_compute_params * p
 void ggml_compute_forward_im2col(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 void ggml_compute_forward_im2col_back_f32(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 void ggml_compute_forward_conv_transpose_2d(const struct ggml_compute_params * params, struct ggml_tensor * dst);
+void ggml_compute_forward_depthwise_conv_2d(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 void ggml_compute_forward_pool_1d(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 void ggml_compute_forward_pool_2d(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 void ggml_compute_forward_pool_2d_back(const struct ggml_compute_params * params, struct ggml_tensor * dst);

src/ggml.c

@@ -956,6 +956,7 @@ static const char * GGML_OP_NAME[GGML_OP_COUNT] = {
     "CONV_TRANSPOSE_1D",
     "IM2COL",
     "IM2COL_BACK",
+    "DEPTHWISE_CONV_2D",
     "CONV_TRANSPOSE_2D",
     "POOL_1D",
     "POOL_2D",
@@ -993,7 +994,7 @@ static const char * GGML_OP_NAME[GGML_OP_COUNT] = {
     "OPT_STEP_ADAMW",
 };
 
-static_assert(GGML_OP_COUNT == 81, "GGML_OP_COUNT != 81");
+static_assert(GGML_OP_COUNT == 82, "GGML_OP_COUNT != 82");
 
 static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = {
     "none",
@@ -1050,6 +1051,7 @@ static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = {
     "conv_transpose_1d(x)",
     "im2col(x)",
     "im2col_back(x)",
+    "depthwise_conv_2d(x)",
     "conv_transpose_2d(x)",
     "pool_1d(x)",
     "pool_2d(x)",
@@ -1087,7 +1089,7 @@ static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = {
     "adamw(x)",
 };
 
-static_assert(GGML_OP_COUNT == 81, "GGML_OP_COUNT != 81");
+static_assert(GGML_OP_COUNT == 82, "GGML_OP_COUNT != 82");
 
 static_assert(GGML_OP_POOL_COUNT == 2, "GGML_OP_POOL_COUNT != 2");
 
@@ -1344,6 +1346,13 @@ bool ggml_is_permuted(const struct ggml_tensor * tensor) {
     return tensor->nb[0] > tensor->nb[1] || tensor->nb[1] > tensor->nb[2] || tensor->nb[2] > tensor->nb[3];
 }
 
+bool ggml_is_contiguous_channels(const struct ggml_tensor * tensor) {
+    return
+        tensor->nb[0] > tensor->nb[2] &&
+        tensor->nb[1] > tensor->nb[0] &&
+        tensor->nb[2] == ggml_type_size(tensor->type);
+}
+
 static inline bool ggml_is_padded_1d(const struct ggml_tensor * tensor) {
     static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function");
 
@@ -4050,6 +4059,44 @@ struct ggml_tensor * ggml_conv_2d_dw(
     return result;
 }
 
+// ggml_depthwise_conv_2d
+
+struct ggml_tensor * ggml_depthwise_conv_2d(
+        struct ggml_context * ctx,
+        struct ggml_tensor  * a,
+        struct ggml_tensor  * b,
+        int                   stride0,
+        int                   stride1,
+        int                   pad0,
+        int                   pad1) {
+    GGML_ASSERT(a->ne[2] == 1);
+    GGML_ASSERT(a->ne[3] == b->ne[2]);
+    int64_t ne[4];
+    ne[0] = ggml_calc_conv_output_size(b->ne[0], a->ne[0], stride0, pad0, 1);
+    ne[1] = ggml_calc_conv_output_size(b->ne[1], a->ne[1], stride1, pad1, 1);
+    ne[2] = b->ne[2];
+    ne[3] = b->ne[3];
+
+    struct ggml_tensor * result = ggml_new_tensor(ctx, b->type, 4, ne);
+
+    if (ggml_is_contiguous_channels(b)) {
+        // Result will be permuted the same way as input (CWHN order)
+        const int64_t type_size = ggml_type_size(result->type);
+        GGML_ASSERT(ggml_blck_size(result->type) == 1);
+        result->nb[0] = result->ne[2] * type_size;
+        result->nb[1] = result->ne[0] * result->nb[0];
+        result->nb[2] = type_size;
+    }
+
+    int32_t params[] = { stride0, stride1, pad0, pad1 };
+    ggml_set_op_params(result, params, sizeof(params));
+
+    result->op     = GGML_OP_DEPTHWISE_CONV_2D;
+    result->src[0] = a;
+    result->src[1] = b;
+    return result;
+}
+
 // ggml_conv_transpose_2d_p0
 
 static int64_t ggml_calc_conv_transpose_output_size(int64_t ins, int64_t ks, int s, int p) {

tests/CMakeLists.txt

@@ -384,6 +384,16 @@ add_test(NAME ${TEST_TARGET} COMMAND $<TARGET_FILE:${TEST_TARGET}>)
 set_property(TEST ${TEST_TARGET} PROPERTY ENVIRONMENT "LLVM_PROFILE_FILE=${TEST_TARGET}.profraw")
 
 
+#
+# test-depthwise-conv2d
+
+set(TEST_TARGET test-depthwise-conv2d)
+add_executable(${TEST_TARGET} ${TEST_TARGET}.cpp)
+target_link_libraries(${TEST_TARGET} PRIVATE ggml)
+add_test(NAME ${TEST_TARGET} COMMAND $<TARGET_FILE:${TEST_TARGET}>)
+set_property(TEST ${TEST_TARGET} PROPERTY ENVIRONMENT "LLVM_PROFILE_FILE=${TEST_TARGET}.profraw")
+
+
 #
 # test-mul-mat