[GitHub] [tvm] rijulg commented on a change in pull request #8006: Fix executor for different compilers

[GitBox <gi...@apache.org>]

rijulg commented on a change in pull request #8006:
URL: https://github.com/apache/tvm/pull/8006#discussion_r628709267



##########
File path: src/runtime/crt/aot_executor/aot_executor.c
##########
@@ -37,29 +37,26 @@ tvm_crt_error_t tvm_runtime_run(const tvm_model_t* model, void** inputs, void**
   TVMValue tvm_values[model->num_input_tensors + model->num_output_tensors];  // NOLINT
   int32_t tvm_typeids[model->num_input_tensors + model->num_output_tensors];  // NOLINT
 
-  for (int i = 0; i < model->num_input_tensors; i++) {
-    tensors[i] = (DLTensor){
-        .device = fake_device,
-        .data = inputs[i],
-        .shape = &fake_shape,
-        .ndim = fake_dims,
-        .byte_offset = 0,
-        .strides = NULL,
-    };
+  for (uint32_t i = 0; i < model->num_input_tensors; i++) {

Review comment:
       I think this is better but we will have to change the type of `model->num_input_tensors` as well.




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org