* [VTA] improved virtual memory mapping

* Update virtual_memory.cc

* [PYTHON][FFI] Cythonize NDArray.copyto

* Cythonize the shape property

* [DOCS] Mention Ninja build system in install/from_source.rst

* Address comments

One of the loops in hybrid_nms used for
performing the max_output_size reordering
was incorrectly designated as parallel
resulting in incorrect behaviour. This patch
changes that loop to a serial loop.

Change-Id: I97184f5887f5f028d8ab339fa2808eb7630a4017

* Add Expand to onnx.py

* add test function for expand

* Fix a onnx frontend test

* Add tests for the value itself instead of shape only on test_expand

* Cleaned up some unnecessary modifications.

* Update legacy places from nnvm to relay.

This PR prepares the current mainline to remove nnvm compiler dep.

* remove legacy stage

Change-Id: If2075df5475505f2da87dae7145af5a7ab83d8a4

* Use the best tuner possible

* Add comment denoting availability of better tuners

* Fix typos and wording

* Replace sigmoid() with tanh() in tests for TANH

* Fix bias_add gradient

A change caused collapse_sum_like to reject implicit dimension
broadcasting for bias_add gradient, so switch to explicit sum reduction
on the non-bias axis dimensions.

* Lint fix

* fix mul rewrite

* register Realize Rewrite for global avg pool and add test

* remove unnecessary check

* improve the test case

* [Fix][Hybrid] Fix hybrid script to support array of tensors

* add test case

* clean up

* trigger ci

* [TOPI] implement pool3d op

* use PoolInferCorrectLayout for both 2d and 3d pooling

* unify MakeMaxPool and MakeAvgPool

* fix serialization precision loss in float

When we want to serialize a tvm.tensor object(like pickle)， we will get a precision loss cause by std::to_string()。
For example, a2.value will be 0.0 while a.value=0.00000001 in the following: 
    import tvm
    import pickle
    a = tvm.const(0.00000001, 'float32')
    a2 = pickle.loads(pickle.dumps(a))

* remove line end spaces

This imports the CUDA schedules to rocm.

There is significant duplication between functions.

Spotted while looking to move the tensorflow and tflite framework support to later than
1.13.1. The tests barf around resize_nearest_neighbour not ignoring the attribute
'helpful_pixel_centers'.

That upgrade is a separate discussion while this can go in
independently.

Thanks,
Ramana

This PR tries to increase TSIM performance by introducing multi-threading support.

* Fix

* clean up

* [Relay][Fix] Fix alter op layout when calling a global var

* add test case