deploy.md 4.02 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
Deploy Compiled Modules
=======================
NNVM compiled modules are fully embedded in TVM runtime as long as ```GRAPH_RUNTIME``` option
is enabled in tvm runtime. Check out the [TVM documentation](http://docs.tvmlang.org/) for
how to deploy TVM runtime to your system.

In a nutshell, we will need three items to deploy a compiled module.
Checkout our tutorials on getting started with NNVM compiler for more details.

- The graph json data which contains the execution graph.
- The tvm module library of compiled functions.
- The parameter blobs for stored parameters.

We can then use TVM's runtime API to deploy the compiled module.
Here is an example in python.

```python
import tvm

# tvm module for compiled functions.
loaded_lib = tvm.module.load("deploy.so")
# json graph
loaded_json = open(temp.relpath("deploy.json")).read()
# parameters in binary
loaded_params = bytearray(open(temp.relpath("deploy.params"), "rb").read())

fcreate = tvm.get_global_func("tvm.graph_runtime.create")
ctx = tvm.gpu(0)
gmodule = fcreate(loaded_json, loaded_lib, ctx.device_type, ctx.device_id)
set_input, get_output, run = gmodule["set_input"], gmodule["get_output"], gmodule["run"]
set_input("x", tvm.nd.array(x_np))
gmodule["load_params"](loaded_params)
run()
out = tvm.nd.empty(shape)
get_output(0, out)
print(out.asnumpy())
```
WolffyChen committed
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52

An example in c++.
```cpp
#include <dlpack/dlpack.h>
#include <tvm/runtime/module.h>
#include <tvm/runtime/registry.h>
#include <tvm/runtime/packed_func.h>

#include <fstream>
#include <iterator>
#include <algorithm>

int main()
{
    // tvm module for compiled functions
53
    tvm::runtime::Module mod_syslib = tvm::runtime::Module::LoadFromFile("deploy.so");
WolffyChen committed
54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69

    // json graph
    std::ifstream json_in("deploy.json", std::ios::in);
    std::string json_data((std::istreambuf_iterator<char>(json_in)), std::istreambuf_iterator<char>());
    json_in.close();

    // parameters in binary
    std::ifstream params_in("deploy.params", std::ios::binary);
    std::string params_data((std::istreambuf_iterator<char>(params_in)), std::istreambuf_iterator<char>());
    params_in.close();

    // parameters need to be TVMByteArray type to indicate the binary data
    TVMByteArray params_arr;
    params_arr.data = params_data.c_str();
    params_arr.size = params_data.length();

WolffyChen committed
70
    int dtype_code = kDLFloat;
WolffyChen committed
71 72
    int dtype_bits = 32;
    int dtype_lanes = 1;
WolffyChen committed
73
    int device_type = kDLCPU;
WolffyChen committed
74 75
    int device_id = 0;

WolffyChen committed
76 77 78
    // get global function module for graph runtime
    tvm::runtime::Module mod = (*tvm::runtime::Registry::Get("tvm.graph_runtime.create"))(json_data, mod_syslib, device_type, device_id);

WolffyChen committed
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
    DLTensor* x;
    int in_ndim = 4;
    int64_t in_shape[4] = {1, 3, 224, 224};
    TVMArrayAlloc(in_shape, in_ndim, dtype_code, dtype_bits, dtype_lanes, device_type, device_id, &x);
    // load image data saved in binary
    std::ifstream data_fin("cat.bin", std::ios::binary);
    data_fin.read(static_cast<char*>(x->data), 3 * 224 * 224 * 4);

    // get the function from the module(set input data)
    tvm::runtime::PackedFunc set_input = mod.GetFunction("set_input");
    set_input("data", x);

    // get the function from the module(load patameters)
    tvm::runtime::PackedFunc load_params = mod.GetFunction("load_params");
    load_params(params_arr);

    // get the function from the module(run it)
    tvm::runtime::PackedFunc run = mod.GetFunction("run");
    run();

    DLTensor* y;
    int out_ndim = 1;
    int64_t out_shape[1] = {1000, };
    TVMArrayAlloc(out_shape, out_ndim, dtype_code, dtype_bits, dtype_lanes, device_type, device_id, &y);

    // get the function from the module(get output data)
    tvm::runtime::PackedFunc get_output = mod.GetFunction("get_output");
    get_output(0, y);

    // get the maximum position in output vector
    auto y_iter = static_cast<float*>(y->data);
    auto max_iter = std::max_element(y_iter, y_iter + 1000);
    auto max_index = std::distance(y_iter, max_iter);
    std::cout << "The maximum position in output vector is: " << max_index << std::endl;

    TVMArrayFree(x);
    TVMArrayFree(y);

    return 0;
}
```