cuda/cpu: add im2col_3d support

ggml/include/ggml.h

@@ -511,6 +511,7 @@ extern "C" {
         GGML_OP_CONV_TRANSPOSE_1D,
         GGML_OP_IM2COL,
         GGML_OP_IM2COL_BACK,
+        GGML_OP_IM2COL_3D,
         GGML_OP_CONV_2D,
         GGML_OP_CONV_3D,
         GGML_OP_CONV_2D_DW,

ggml/src/ggml-cpu/ggml-cpu.c

@@ -1876,6 +1876,10 @@ static void ggml_compute_forward(struct ggml_compute_params * params, struct ggm
             {
                 ggml_compute_forward_im2col_back_f32(params, tensor);
             } break;
+        case GGML_OP_IM2COL_3D:
+            {
+                ggml_compute_forward_im2col_3d(params, tensor);
+            } break;
         case GGML_OP_CONV_2D:
             {
                 ggml_compute_forward_conv_2d(params, tensor);
@@ -2255,6 +2259,7 @@ 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_IM2COL_3D:
         case GGML_OP_CONV_2D:
         case GGML_OP_CONV_3D:
         case GGML_OP_CONV_2D_DW:

ggml/src/ggml-cpu/ops.cpp

@@ -7027,6 +7027,209 @@ void ggml_compute_forward_im2col_back_f32(
     }
 }
 
+
+// ggml_compute_forward_im2col_3d_f16
+// src0: kernel [OC*IC, KD, KH, KW]
+// src1: image [N*IC, ID, IH, IW]
+// dst:  result [N*OD, OH, OW, IC * KD * KH * KW]
+static void ggml_compute_forward_im2col_3d_f16(
+        const ggml_compute_params * params,
+              ggml_tensor * dst) {
+
+    const ggml_tensor * src0 = dst->src[0];
+    const ggml_tensor * src1 = dst->src[1];
+
+    GGML_ASSERT(src0->type == GGML_TYPE_F16);
+    GGML_ASSERT(src1->type == GGML_TYPE_F32);
+    GGML_ASSERT( dst->type == GGML_TYPE_F16);
+
+    GGML_TENSOR_BINARY_OP_LOCALS;
+
+    const int32_t s0 = ((const int32_t *)(dst->op_params))[0];
+    const int32_t s1 = ((const int32_t *)(dst->op_params))[1];
+    const int32_t s2 = ((const int32_t *)(dst->op_params))[2];
+    const int32_t p0 = ((const int32_t *)(dst->op_params))[3];
+    const int32_t p1 = ((const int32_t *)(dst->op_params))[4];
+    const int32_t p2 = ((const int32_t *)(dst->op_params))[5];
+    const int32_t d0 = ((const int32_t *)(dst->op_params))[6];
+    const int32_t d1 = ((const int32_t *)(dst->op_params))[7];
+    const int32_t d2 = ((const int32_t *)(dst->op_params))[8];
+    const int32_t IC = ((const int32_t *)(dst->op_params))[9];
+
+
+    const int ith = params->ith;
+    const int nth = params->nth;
+
+    const int64_t N  = ne13 / IC;
+    const int64_t ID = ne12;
+    const int64_t IH = ne11;
+    const int64_t IW = ne10;
+
+    const int64_t OC = ne03 / IC;
+    const int64_t KD = ne02;
+    const int64_t KH = ne01;
+    const int64_t KW = ne00;
+
+    const int64_t OD = ne3 / N;
+    const int64_t OH = ne2;
+    const int64_t OW = ne1;
+    const int64_t OH_OW = OH*OW;
+    const int64_t KD_KH_KW = KD*KH*KW;
+    const int64_t ID_IH_IW = ID*IH*IW;
+    const int64_t KH_KW = KH*KW;
+    const int64_t IH_IW = IH*IW;
+    const int64_t IC_KD_KH_KW = IC*KD*KH*KW;
+
+    GGML_ASSERT(nb10 == sizeof(float));
+
+    // im2col: [N*IC, ID, IH, IW] => [N*OD, OH, OW, IC * KD * KH * KW]
+    {
+        ggml_fp16_t * const wdata = (ggml_fp16_t *) dst->data;
+
+        for (int64_t in = 0; in < N; in++) {
+            for (int64_t iod = 0; iod < OD; iod++) {
+                for (int64_t ioh = 0; ioh < OH; ioh++) {
+                    for (int64_t iow = 0; iow < OW; iow++) {
+                        for (int64_t iic = ith; iic < IC; iic += nth) {
+
+                            // micro kernel
+                            ggml_fp16_t * dst_data = wdata + (in*OH_OW + ioh*OW + iow)*IC_KD_KH_KW; // [IC, KD, KH, KW]
+                            const float * const src_data = (float *) src1->data + (in*IC + iic)*ID_IH_IW; // [ID, IH, IW]
+
+                            for (int64_t ikd = 0; ikd < KD; ikd++) {
+                                for (int64_t ikh = 0; ikh < KH; ikh++) {
+                                    for (int64_t ikw = 0; ikw < KW; ikw++) {
+                                        const int64_t iiw = iow*s0 + ikw*d0 - p0;
+                                        const int64_t iih = ioh*s1 + ikh*d1 - p1;
+                                        const int64_t iid = iod*s2 + ikd*d2 - p2;
+
+                                        if (iid < 0 || iid >= ID || iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) {
+                                            dst_data[iic*KD_KH_KW + ikd * KH_KW + ikh*KW + ikw] = 0;
+                                        } else {
+                                            dst_data[iic*KD_KH_KW + ikd * KH_KW + ikh*KW + ikw] = GGML_CPU_FP32_TO_FP16(src_data[iid*IH_IW + iih*IW + iiw]);
+                                        }
+                                    }
+                                }
+                            }
+                        }
+                    }
+                }
+            }
+        }
+    }
+}
+
+// ggml_compute_forward_im2col_3d_f32
+// src0: kernel [OC*IC, KD, KH, KW]
+// src1: image [N*IC, ID, IH, IW]
+// dst:  result [N*OD, OH, OW, IC * KD * KH * KW]
+static void ggml_compute_forward_im2col_3d_f32(
+        const ggml_compute_params * params,
+              ggml_tensor * dst) {
+
+    const ggml_tensor * src0 = dst->src[0];
+    const ggml_tensor * src1 = dst->src[1];
+
+    GGML_ASSERT(src1->type == GGML_TYPE_F32);
+    GGML_ASSERT( dst->type == GGML_TYPE_F32);
+
+    GGML_TENSOR_BINARY_OP_LOCALS;
+
+    const int32_t s0 = ((const int32_t *)(dst->op_params))[0];
+    const int32_t s1 = ((const int32_t *)(dst->op_params))[1];
+    const int32_t s2 = ((const int32_t *)(dst->op_params))[2];
+    const int32_t p0 = ((const int32_t *)(dst->op_params))[3];
+    const int32_t p1 = ((const int32_t *)(dst->op_params))[4];
+    const int32_t p2 = ((const int32_t *)(dst->op_params))[5];
+    const int32_t d0 = ((const int32_t *)(dst->op_params))[6];
+    const int32_t d1 = ((const int32_t *)(dst->op_params))[7];
+    const int32_t d2 = ((const int32_t *)(dst->op_params))[8];
+    const int32_t IC = ((const int32_t *)(dst->op_params))[9];
+
+
+    const int ith = params->ith;
+    const int nth = params->nth;
+
+    const int64_t N  = ne13 / IC;
+    const int64_t ID = ne12;
+    const int64_t IH = ne11;
+    const int64_t IW = ne10;
+
+    const int64_t OC = ne03 / IC;
+    const int64_t KD = ne02;
+    const int64_t KH = ne01;
+    const int64_t KW = ne00;
+
+    const int64_t OD = ne3 / N;
+    const int64_t OH = ne2;
+    const int64_t OW = ne1;
+
+    const int64_t OH_OW = OH*OW;
+    const int64_t KD_KH_KW = KD*KH*KW;
+    const int64_t ID_IH_IW = ID*IH*IW;
+    const int64_t KH_KW = KH*KW;
+    const int64_t IH_IW = IH*IW;
+    const int64_t IC_KD_KH_KW = IC*KD*KH*KW;
+
+    GGML_ASSERT(nb10 == sizeof(float));
+
+    // im2col: [N*IC, ID, IH, IW] => [N*OD, OH, OW, IC * KD * KH * KW]
+    {
+        float * const wdata = (float *) dst->data;
+
+        for (int64_t in = 0; in < N; in++) {
+            for (int64_t iod = 0; iod < OD; iod++) {
+                for (int64_t ioh = 0; ioh < OH; ioh++) {
+                    for (int64_t iow = 0; iow < OW; iow++) {
+                        for (int64_t iic = ith; iic < IC; iic += nth) {
+
+                            // micro kernel
+                            float * dst_data = wdata + (in*OH_OW + ioh*OW + iow)*IC_KD_KH_KW; // [IC, KD, KH, KW]
+                            const float * const src_data = (float *) src1->data + (in*IC + iic)*ID_IH_IW; // [ID, IH, IW]
+
+                            for (int64_t ikd = 0; ikd < KD; ikd++) {
+                                for (int64_t ikh = 0; ikh < KH; ikh++) {
+                                    for (int64_t ikw = 0; ikw < KW; ikw++) {
+                                        const int64_t iiw = iow*s0 + ikw*d0 - p0;
+                                        const int64_t iih = ioh*s1 + ikh*d1 - p1;
+                                        const int64_t iid = iod*s2 + ikd*d2 - p2;
+
+                                        if (iid < 0 || iid >= ID || iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) {
+                                            dst_data[iic*KD_KH_KW + ikd * KH_KW + ikh*KW + ikw] = 0;
+                                        } else {
+                                            dst_data[iic*KD_KH_KW + ikd * KH_KW + ikh*KW + ikw] = src_data[iid*IH_IW + iih*IW + iiw];
+                                        }
+                                    }
+                                }
+                            }
+                        }
+                    }
+                }
+            }
+        }
+    }
+}
+
+
+void ggml_compute_forward_im2col_3d(
+        const ggml_compute_params * params,
+              ggml_tensor * dst) {
+    switch (dst->type) {
+        case GGML_TYPE_F16:
+            {
+                ggml_compute_forward_im2col_3d_f16(params, dst);
+            } break;
+        case GGML_TYPE_F32:
+            {
+                ggml_compute_forward_im2col_3d_f32(params, dst);
+            } break;
+        default:
+            {
+                GGML_ABORT("fatal error");
+            }
+    }
+}
+
 static void ggml_call_mul_mat(ggml_type type, const ggml_compute_params * params, int64_t m, int64_t n, int64_t k,
                               void * a, void * b, float * c) {
     const ggml_type_traits * traits = ggml_get_type_traits(type);

ggml/src/ggml-cpu/ops.h

@@ -69,6 +69,7 @@ void ggml_compute_forward_clamp(const struct ggml_compute_params * params, struc
 void ggml_compute_forward_conv_transpose_1d(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 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_im2col_3d(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 void ggml_compute_forward_conv_2d(const struct ggml_compute_params * params, struct ggml_tensor * dst);
 void ggml_compute_forward_conv_3d(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);

ggml/src/ggml-cuda/ggml-cuda.cu

@@ -2452,6 +2452,9 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg
         case GGML_OP_IM2COL:
             ggml_cuda_op_im2col(ctx, dst);
             break;
+        case GGML_OP_IM2COL_3D:
+            ggml_cuda_op_im2col_3d(ctx, dst);
+            break;
         case GGML_OP_CONV_2D:
             ggml_cuda_op_conv2d(ctx, dst);
             break;
@@ -3559,6 +3562,7 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
             return op->src[0]->nb[0] == ggml_type_size(op->src[0]->type) && ggml_is_contiguous_2(op->src[0]);
         }
         case GGML_OP_IM2COL:
+        case GGML_OP_IM2COL_3D:
         case GGML_OP_CONV_2D:
         case GGML_OP_CONV_2D_DW:
         case GGML_OP_CONV_TRANSPOSE_2D:

ggml/src/ggml-cuda/im2col.cu

@@ -112,3 +112,125 @@ void ggml_cuda_op_im2col(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
         im2col_cuda_f32(src1_d, (float *) dst_d, IW, IH, OW, OH, KW, KH, IC, N, IC_IH_IW, IH_IW, s0, s1, p0, p1, d0, d1, stream);
     }
 }
+
+// [N*IC, ID, IH, IW] => [N*OD, OH, OW, IC * KD * KH * KW]
+template <typename T>
+static  __global__ void im2col_3d_kernel(
+        const float * src, T * dst,
+        int64_t N, int64_t IC, int64_t ID, int64_t IH, int64_t IW, int64_t OC,
+        int64_t KD, int64_t KH, int64_t KW, int64_t OD, int64_t OH, int64_t OW,
+        int64_t OH_OW, int64_t KD_KH_KW, int64_t ID_IH_IW, int64_t KH_KW, int64_t IH_IW,
+        int64_t IC_KD_KH_KW, int64_t OW_KD_KH_KW, int64_t N_OD_IC, int64_t OD_IC,
+        int s0, int s1, int s2, int p0, int p1, int p2, int d0, int d1, int d2) {
+    const int64_t i = threadIdx.x + blockIdx.x * blockDim.x;
+    if (i >= OW_KD_KH_KW) {
+        return;
+    }
+
+    const int64_t iow = i / KD_KH_KW;
+    const int64_t ikd = (i - iow * KD_KH_KW) / KH_KW;
+    const int64_t ikh = (i - iow * KD_KH_KW - ikd * KH_KW) / KW;
+    const int64_t ikw = i % KW;
+
+    const int64_t  ioh = blockIdx.y;
+    for (int64_t iz = blockIdx.z; iz < N_OD_IC; iz+=MAX_GRID_DIM_Z) {
+        const int64_t in  = iz / OD_IC;
+        const int64_t iod = (iz - in*OD_IC) / IC;
+        const int64_t iic = iz % IC;
+
+        const int64_t iiw = iow * s0 + ikw * d0 - p0;
+        const int64_t iih = ioh * s1 + ikh * d1 - p1;
+        const int64_t iid = iod * s2 + ikd * d2 - p2;
+
+        const int64_t offset_dst = (in*OH_OW + ioh*OW + iow)*IC_KD_KH_KW + iic*KD_KH_KW + ikd * KH_KW + ikh*KW + ikw;
+
+        if (iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) {
+            dst[offset_dst] = 0.0f;
+        } else {
+            const int64_t offset_src = (in*IC + iic)*ID_IH_IW + iid*IH_IW + iih*IW + iiw;
+            dst[offset_dst] = src[offset_src];
+        }
+    }
+}
+
+// [N*IC, ID, IH, IW] => [N*OD, OH, OW, IC * KD * KH * KW]
+template <typename T>
+static void im2col_3d_cuda(const float * src, T* dst,
+    int64_t N, int64_t IC, int64_t ID, int64_t IH, int64_t IW, int64_t OC,
+    int64_t KD, int64_t KH, int64_t KW, int64_t OD, int64_t OH, int64_t OW,
+    int s0, int s1, int s2, int p0, int p1, int p2, int d0, int d1, int d2, cudaStream_t stream) {
+    const int64_t OH_OW = OH*OW;
+    const int64_t KD_KH_KW = KD*KH*KW;
+    const int64_t ID_IH_IW = ID*IH*IW;
+    const int64_t KH_KW = KH*KW;
+    const int64_t IH_IW = IH*IW;
+    const int64_t IC_KD_KH_KW = IC*KD*KH*KW;
+    const int64_t OW_KD_KH_KW = OW*KD*KH*KW;
+    const int64_t N_OD_IC = N*OD*IC;
+    const int64_t OD_IC = OD*IC;
+    const int64_t num_blocks = (OW_KD_KH_KW + CUDA_IM2COL_BLOCK_SIZE - 1) / CUDA_IM2COL_BLOCK_SIZE;
+    dim3 block_nums(num_blocks, OH, MIN(N_OD_IC, MAX_GRID_DIM_Z));
+    im2col_3d_kernel<<<block_nums, MIN(OW_KD_KH_KW, CUDA_IM2COL_BLOCK_SIZE) , 0, stream>>>(src, dst, N, IC, ID, IH, IW, OC, KD, KH, KW, OD, OH, OW,
+                                                                                           OH_OW, KD_KH_KW, ID_IH_IW, KH_KW, IH_IW, IC_KD_KH_KW, 
+                                                                                           OW_KD_KH_KW, N_OD_IC, OD_IC, s0, s1, s2, p0, p1, p2, d0, d1, d2);
+}
+
+static void im2col_3d_cuda_f16(const float * src, half * dst,
+    int64_t N, int64_t IC, int64_t ID, int64_t IH, int64_t IW, int64_t OC,
+    int64_t KD, int64_t KH, int64_t KW, int64_t OD, int64_t OH, int64_t OW,
+    int s0, int s1, int s2, int p0, int p1, int p2, int d0, int d1, int d2, cudaStream_t stream) {
+
+    im2col_3d_cuda<half>(src, dst, N, IC, ID, IH, IW, OC, KD, KH, KW, OD, OH, OW, s0, s1, s2, p0, p1, p2, d0, d1, d2, stream);
+}
+
+static void im2col_3d_cuda_f32(const float * src, float * dst,
+    int64_t N, int64_t IC, int64_t ID, int64_t IH, int64_t IW, int64_t OC,
+    int64_t KD, int64_t KH, int64_t KW, int64_t OD, int64_t OH, int64_t OW,
+    int s0, int s1, int s2, int p0, int p1, int p2, int d0, int d1, int d2, cudaStream_t stream) {
+
+    im2col_3d_cuda<float>(src, dst, N, IC, ID, IH, IW, OC, KD, KH, KW, OD, OH, OW, s0, s1, s2, p0, p1, p2, d0, d1, d2, stream);
+}
+
+void ggml_cuda_op_im2col_3d(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
+    const ggml_tensor * src0 = dst->src[0];
+    const ggml_tensor * src1 = dst->src[1];
+    const float * src1_d = (const float *)src1->data;
+    float * dst_d = (float *)dst->data;
+    cudaStream_t stream = ctx.stream();
+
+    GGML_ASSERT(src1->type == GGML_TYPE_F32);
+    GGML_ASSERT( dst->type == GGML_TYPE_F16 || dst->type == GGML_TYPE_F32);
+
+    GGML_TENSOR_BINARY_OP_LOCALS
+
+    const int32_t s0 = ((const int32_t *)(dst->op_params))[0];
+    const int32_t s1 = ((const int32_t *)(dst->op_params))[1];
+    const int32_t s2 = ((const int32_t *)(dst->op_params))[2];
+    const int32_t p0 = ((const int32_t *)(dst->op_params))[3];
+    const int32_t p1 = ((const int32_t *)(dst->op_params))[4];
+    const int32_t p2 = ((const int32_t *)(dst->op_params))[5];
+    const int32_t d0 = ((const int32_t *)(dst->op_params))[6];
+    const int32_t d1 = ((const int32_t *)(dst->op_params))[7];
+    const int32_t d2 = ((const int32_t *)(dst->op_params))[8];
+    const int32_t IC = ((const int32_t *)(dst->op_params))[9];
+
+    const int64_t N  = ne13 / IC;
+    const int64_t ID = ne12;
+    const int64_t IH = ne11;
+    const int64_t IW = ne10;
+
+    const int64_t OC = ne03 / IC;
+    const int64_t KD = ne02;
+    const int64_t KH = ne01;
+    const int64_t KW = ne00;
+
+    const int64_t OD = ne3 / N;
+    const int64_t OH = ne2;
+    const int64_t OW = ne1;
+
+    if(dst->type == GGML_TYPE_F16) {
+        im2col_3d_cuda_f16(src1_d, (half *) dst_d, N, IC, ID, IH, IW, OC, KD, KH, KW, OD, OH, OW, s0, s1, s2, p0, p1, p2, d0, d1, d2, stream);
+    } else {
+        im2col_3d_cuda_f32(src1_d, (float *) dst_d, N, IC, ID, IH, IW, OC, KD, KH, KW, OD, OH, OW, s0, s1, s2, p0, p1, p2, d0, d1, d2, stream);
+    }
+}

ggml/src/ggml-cuda/im2col.cuh

@@ -3,3 +3,4 @@
 #define CUDA_IM2COL_BLOCK_SIZE 256
 
 void ggml_cuda_op_im2col(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
+void ggml_cuda_op_im2col_3d(ggml_backend_cuda_context & ctx, ggml_tensor * dst);

ggml/src/ggml.c

@@ -974,6 +974,7 @@ static const char * GGML_OP_NAME[GGML_OP_COUNT] = {
     "CONV_TRANSPOSE_1D",
     "IM2COL",
     "IM2COL_BACK",
+    "IM2COL_3D",
     "CONV_2D",
     "CONV_3D",
     "CONV_2D_DW",
@@ -1077,6 +1078,7 @@ static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = {
     "conv_transpose_1d(x)",
     "im2col(x)",
     "im2col_back(x)",
+    "im2col_3d(x)"
     "conv_2d(x)",
     "conv_3d(x)",
     "conv_2d_dw(x)",
@@ -4379,7 +4381,6 @@ struct ggml_tensor * ggml_im2col_3d(
         int                   d1, // dilation height
         int                   d2, // dilation depth
         enum ggml_type        dst_type) {
-
     const int64_t N = b->ne[3] / IC;
     const int64_t ID = b->ne[2];
     const int64_t IH = b->ne[1];
@@ -4390,22 +4391,25 @@ struct ggml_tensor * ggml_im2col_3d(
     const int64_t KH = a->ne[1];
     const int64_t KW = a->ne[0];
     const int64_t OD = ggml_calc_conv_output_size(ID, KD, s2, p2, d2);
+    const int64_t OH = ggml_calc_conv_output_size(IH, KH, s1, p1, d1);
+    const int64_t OW = ggml_calc_conv_output_size(IW, KW, s0, p0, d0);
 
-    struct ggml_tensor* img = ggml_reshape_4d(ctx, b, IW*IH, ID, IC, N); // [N, IC, ID, IH * IW]
-    img = ggml_cont(ctx, ggml_permute(ctx, img, 2, 0, 1, 3)); // [N, IH*IW, IC, ID]
-    img = ggml_reshape_3d(ctx, b, ID, IC, IW*IH*N); // [N*IH*IW, IC, ID]
+    GGML_ASSERT((OD > 0)  && "b too small compared to a");
+    GGML_ASSERT((OH > 0)  && "b too small compared to a");
+    GGML_ASSERT((OW > 0)  && "b too small compared to a");
 
-    a = ggml_reshape_3d(ctx, a, KD, IC, OC*KW*KH); // [OC*KW*KH, IC, KD]
-    img = ggml_im2col(ctx, a, img, s2, 1, p2, 0, d2, 1, false, GGML_TYPE_F32); // [N*IH*IW, OD, IC*KD]
 
-    img = ggml_reshape_4d(ctx, img, IC*KD, OD, IW*IH, N); // [N, IH*IW, OD, IC*KD]
-    img = ggml_cont(ctx, ggml_permute(ctx, img, 1, 2, 0, 3)); // [N, OD, IC*KD, IH*IW]
-    img = ggml_reshape_4d(ctx, img, IW, IH, IC*KD, OD*N); // [N*OD, IC*KD, IH, IW]
+    const int64_t ne[4] = {KW*KH*KD*IC, OW, OH, OD*N};
 
-    a = ggml_reshape_4d(ctx, a, KW, KH, IC*KD, OC); // [OC, KD*IC, KH, KW]
+    struct ggml_tensor * result = ggml_new_tensor(ctx, dst_type, 4, ne);
+    int32_t params[] = { s0, s1, s2, p0, p1, p2, d0, d1, d2, (int32_t)IC};
+    ggml_set_op_params(result, params, sizeof(params));
 
-    img = ggml_im2col(ctx, a, img, s0, s1, p0, p1, d0, d1, true, dst_type); // [N * OD, OH, OW, IC * KD * KH * KW]
-    return img;
+    result->op     = GGML_OP_IM2COL_3D;
+    result->src[0] = a;
+    result->src[1] = b;
+
+    return result;
 }
 
 // a: [OC*IC, KD, KH, KW]