Add UPSCALE

Author: wangshuai09

ggml-cann.cpp

@@ -400,7 +400,8 @@ static bool ggml_cann_compute_forward(ggml_backend_cann_context& ctx,
             break;
         // TODO: Format need NC1HWC0.
         case GGML_OP_UPSCALE:
-            return false;
+            ggml_cann_upsample_nearest2d(ctx, dst);
+            break;
         case GGML_OP_PAD:
             ggml_cann_pad(ctx, dst);
             break;
@@ -687,7 +688,7 @@ GGML_CALL static bool ggml_backend_cann_supports_op(ggml_backend_t backend,
         case GGML_OP_GROUP_NORM:
             return true;
         case GGML_OP_UPSCALE:
-            return false;
+            return true;
         case GGML_OP_PAD:
         case GGML_OP_ARANGE:
             return true;

ggml-cann/aclnn_ops.cpp

@@ -5,6 +5,7 @@
 #include <aclnnop/aclnn_layer_norm.h>
 #include <aclnnop/aclnn_repeat.h>
 #include <aclnnop/aclnn_softmax.h>
+#include <aclnnop/aclnn_upsample_nearest_2d.h>
 #include <aclnnop/aclnn_reduce_sum.h>
 
 #include <cmath>
@@ -486,10 +487,6 @@ void ggml_cann_sum_rows(ggml_backend_cann_context& ctx, ggml_tensor* dst) {
     GGML_ASSERT(dst->ne[0] == 1);
     aclTensor* acl_dst = create_acl_tensor(dst);
 
-    uint64_t workspaceSize = 0;
-    aclOpExecutor* executor;
-    void* workspaceAddr = nullptr;
-
     int64_t reduce_dims_host[] = {3};
     aclIntArray* reduce_dims = aclCreateIntArray(reduce_dims_host, 1);
 
@@ -503,6 +500,41 @@ void ggml_cann_sum_rows(ggml_backend_cann_context& ctx, ggml_tensor* dst) {
     aclrtStream stream = ctx.stream();
     ACL_CHECK(aclnnReduceSum(workspaceAddr, workspaceSize, executor, stream));
 
+    ACL_CHECK(aclDestroyTensor(acl_src));
+    ACL_CHECK(aclDestroyTensor(acl_dst));
+}
+
+void ggml_cann_upsample_nearest2d(ggml_backend_cann_context& ctx, 
+    ggml_tensor* dst) {
+
+    ggml_tensor* src = dst->src[0];
+
+    aclTensor* acl_src = create_acl_tensor(src, nullptr, nullptr, 0, 
+                                           ACL_FORMAT_NCHW);
+    aclTensor* acl_dst = create_acl_tensor(dst, nullptr, nullptr, 0, 
+                                           ACL_FORMAT_NCHW);
+
+    const int scale_factor = dst->op_params[0];
+    std::vector<int64_t> output_size{dst->ne[1], dst->ne[0]};
+    auto output_size_array = aclCreateIntArray(output_size.data(), 2);
+    
+    uint64_t workspaceSize = 0;
+    aclOpExecutor* executor;
+    void* workspaceAddr = nullptr;
+
+    aclrtStream stream = ctx.stream();
+
+    ACL_CHECK(aclnnUpsampleNearest2dGetWorkspaceSize(acl_src, output_size_array, 
+                                                     acl_dst, &workspaceSize, 
+                                                     &executor));
+    if (workspaceSize > 0) {
+        workspaceAddr = ctx.alloc_buffer(workspaceSize);
+    }
+    
+    ACL_CHECK(aclnnUpsampleNearest2d(workspaceAddr, workspaceSize, executor, 
+                                     stream));
+    
+    ACL_CHECK(aclDestroyIntArray(output_size_array));
     ACL_CHECK(aclDestroyTensor(acl_src));
     ACL_CHECK(aclDestroyTensor(acl_dst));
 }

ggml-cann/aclnn_ops.h

@@ -149,4 +149,6 @@ void ggml_cann_activation(ggml_backend_cann_context& ctx, ggml_tensor* dst) {
     ACL_CHECK(aclDestroyTensor(acl_dst));
 }
 
+void ggml_cann_upsample_nearest2d(ggml_backend_cann_context& ctx, ggml_tensor* dst);
+
 #endif  // CANN_ACLNN_OPS