TVM darknet yolov3算子优化与量化代码的配置方法

 TVM darknet yolov3算子优化与量化代码的配置方法

使用以下接口函数

l  tvm.relay.optimize

l  quantize.quantize

实际代码：

 

# convert nnvm to relay

print("convert nnvm symbols into relay function...")

#from nnvm.to_relay import to_relay

func, params = to_relay(sym, shape, 'float32', params=params)

# optimization

print("optimize relay graph...")

with tvm.relay.build_config(opt_level=2):

    func = tvm.relay.optimize(func, target, params)

# quantize

print("apply quantization...")

from tvm.relay import quantize

with quantize.qconfig():

   func = quantize.quantize(func, params)

 

参考链接：

https://github.com/makihiro/tvm_yolov3_sample/blob/master/yolov3_quantize_sample.py

 

完全代码如下

早期版本，可以使用新的TVM版本修改。

 

 

# Licensed to the Apache Software Foundation (ASF) under one

# or more contributor license agreements.  See the NOTICE file

# distributed with this work for additional information

# regarding copyright ownership.  The ASF licenses this file

# to you under the Apache License, Version 2.0 (the

# "License"); you may not use this file except in compliance

# with the License.  You may obtain a copy of the License at

#

#   http://www.apache.org/licenses/LICENSE-2.0

#

# Unless required by applicable law or agreed to in writing,

# software distributed under the License is distributed on an

# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY

# KIND, either express or implied.  See the License for the

# specific language governing permissions and limitations

# under the License.

"""

Compile YOLO-V2 and YOLO-V3 in DarkNet Models

=============================================

**Author**: `Siju Samuel <https://siju-samuel.github.io/>`_

 

This article is an introductory tutorial to deploy darknet models with TVM.

All the required models and libraries will be downloaded from the internet by the script.

This script runs the YOLO-V2 and YOLO-V3 Model with the bounding boxes

Darknet parsing have dependancy with CFFI and CV2 library

Please install CFFI and CV2 before executing this script

 

.. code-block:: bash

 

  pip install cffi

  pip install opencv-python

"""

 

# numpy and matplotlib

import numpy as np

import matplotlib.pyplot as plt

import sys

 

# tvm, relay

import tvm

from tvm import te

from tvm import relay

from ctypes import *

from tvm.contrib.download import download_testdata

from tvm.relay.testing.darknet import __darknetffi__

import tvm.relay.testing.yolo_detection

import tvm.relay.testing.darknet

 

######################################################################

# Choose the model

# -----------------------

# Models are: 'yolov2', 'yolov3' or 'yolov3-tiny'

 

# Model name

MODEL_NAME = "yolov3"

 

######################################################################

# Download required files

# -----------------------

# Download cfg and weights file if first time.

CFG_NAME = MODEL_NAME + ".cfg"

WEIGHTS_NAME = MODEL_NAME + ".weights"

REPO_URL = "https://github.com/dmlc/web-data/blob/main/darknet/"

CFG_URL = REPO_URL + "cfg/" + CFG_NAME + "?raw=true"

WEIGHTS_URL = "https://pjreddie.com/media/files/" + WEIGHTS_NAME

 

cfg_path = download_testdata(CFG_URL, CFG_NAME, module="darknet")

weights_path = download_testdata(WEIGHTS_URL, WEIGHTS_NAME, module="darknet")

 

# Download and Load darknet library

if sys.platform in ["linux", "linux2"]:

    DARKNET_LIB = "libdarknet2.0.so"

    DARKNET_URL = REPO_URL + "lib/" + DARKNET_LIB + "?raw=true"

elif sys.platform == "darwin":

    DARKNET_LIB = "libdarknet_mac2.0.so"

    DARKNET_URL = REPO_URL + "lib_osx/" + DARKNET_LIB + "?raw=true"

else:

    err = "Darknet lib is not supported on {} platform".format(sys.platform)

    raise NotImplementedError(err)

 

lib_path = download_testdata(DARKNET_URL, DARKNET_LIB, module="darknet")

 

DARKNET_LIB = __darknetffi__.dlopen(lib_path)

net = DARKNET_LIB.load_network(cfg_path.encode("utf-8"), weights_path.encode("utf-8"), 0)

dtype = "float32"

batch_size = 1

 

data = np.empty([batch_size, net.c, net.h, net.w], dtype)

shape_dict = {"data": data.shape}

print("Converting darknet to relay functions...")

mod, params = relay.frontend.from_darknet(net, dtype=dtype, shape=data.shape)

######################################################################

# Compile the model on NNVM

# -------------------------

# compile the model

local = True

 

if local:

    target = 'llvm'

    ctx = tvm.cpu(0)

else:

    target = 'cuda'

    ctx = tvm.gpu(0)

 

data = np.empty([batch_size, net.c, net.h, net.w], dtype)

shape = {'data': data.shape}

 

dtype_dict = {}

 

# convert nnvm to relay

print("convert nnvm symbols into relay function...")

#from nnvm.to_relay import to_relay

func, params = to_relay(sym, shape, 'float32', params=params)

# optimization

print("optimize relay graph...")

with tvm.relay.build_config(opt_level=2):

    func = tvm.relay.optimize(func, target, params)

# quantize

print("apply quantization...")

from tvm.relay import quantize

with quantize.qconfig():

   func = quantize.quantize(func, params)

 

# Relay build

print("Compiling the model...")

print(func.astext(show_meta_data=False))

with tvm.relay.build_config(opt_level=3):

    graph, lib, params = tvm.relay.build(func, target=target, params=params)

 

# Save the model

tmp = util.tempdir()

lib_fname = tmp.relpath('model.tar')

lib.export_library(lib_fname)

 

# NNVM

# with nnvm.compiler.build_config(opt_level=2):

#     graph, lib, params = nnvm.compiler.build(sym, target, shape, dtype_dict, params)

 

 

#[neth, netw] = shape['data'][2:]  # Current image shape is 608x608

######################################################################

 

######################################################################

# Import the graph to Relay

# -------------------------

# compile the model

target = tvm.target.Target("llvm", host="llvm")

dev = tvm.cpu(0)

data = np.empty([batch_size, net.c, net.h, net.w], dtype)

shape = {"data": data.shape}

print("Compiling the model...")

with tvm.transform.PassContext(opt_level=3):

    lib = relay.build(mod, target=target, params=params)

 

[neth, netw] = shape["data"][2:]  # Current image shape is 608x608

######################################################################

# Load a test image

# -----------------

test_image = "dog.jpg"

print("Loading the test image...")

img_url = REPO_URL + "data/" + test_image + "?raw=true"

img_path = download_testdata(img_url, test_image, "data")

 

data = tvm.relay.testing.darknet.load_image(img_path, netw, neth)

######################################################################

# Execute on TVM Runtime

# ----------------------

# The process is no different from other examples.

from tvm.contrib import graph_executor

 

m = graph_executor.GraphModule(lib["default"](dev))

 

# set inputs

m.set_input("data", tvm.nd.array(data.astype(dtype)))

# execute

print("Running the test image...")

 

# detection

# thresholds

thresh = 0.5

nms_thresh = 0.45

 

m.run()

# get outputs

tvm_out = []

if MODEL_NAME == "yolov2":

    layer_out = {}

    layer_out["type"] = "Region"

    # Get the region layer attributes (n, out_c, out_h, out_w, classes, coords, background)

    layer_attr = m.get_output(2).numpy()

    layer_out["biases"] = m.get_output(1).numpy()

    out_shape = (layer_attr[0], layer_attr[1] // layer_attr[0], layer_attr[2], layer_attr[3])

    layer_out["output"] = m.get_output(0).numpy().reshape(out_shape)

    layer_out["classes"] = layer_attr[4]

    layer_out["coords"] = layer_attr[5]

    layer_out["background"] = layer_attr[6]

    tvm_out.append(layer_out)

 

elif MODEL_NAME == "yolov3":

    for i in range(3):

        layer_out = {}

        layer_out["type"] = "Yolo"

        # Get the yolo layer attributes (n, out_c, out_h, out_w, classes, total)

        layer_attr = m.get_output(i * 4 + 3).numpy()

        layer_out["biases"] = m.get_output(i * 4 + 2).numpy()

        layer_out["mask"] = m.get_output(i * 4 + 1).numpy()

        out_shape = (layer_attr[0], layer_attr[1] // layer_attr[0], layer_attr[2], layer_attr[3])

        layer_out["output"] = m.get_output(i * 4).numpy().reshape(out_shape)

        layer_out["classes"] = layer_attr[4]

        tvm_out.append(layer_out)

 

elif MODEL_NAME == "yolov3-tiny":

    for i in range(2):

        layer_out = {}

        layer_out["type"] = "Yolo"

        # Get the yolo layer attributes (n, out_c, out_h, out_w, classes, total)

        layer_attr = m.get_output(i * 4 + 3).numpy()

        layer_out["biases"] = m.get_output(i * 4 + 2).numpy()

        layer_out["mask"] = m.get_output(i * 4 + 1).numpy()

        out_shape = (layer_attr[0], layer_attr[1] // layer_attr[0], layer_attr[2], layer_attr[3])

        layer_out["output"] = m.get_output(i * 4).numpy().reshape(out_shape)

        layer_out["classes"] = layer_attr[4]

        tvm_out.append(layer_out)

        thresh = 0.560

 

# do the detection and bring up the bounding boxes

img = tvm.relay.testing.darknet.load_image_color(img_path)

_, im_h, im_w = img.shape

dets = tvm.relay.testing.yolo_detection.fill_network_boxes(

    (netw, neth), (im_w, im_h), thresh, 1, tvm_out

)

last_layer = net.layers[net.n - 1]

tvm.relay.testing.yolo_detection.do_nms_sort(dets, last_layer.classes, nms_thresh)

 

coco_name = "coco.names"

coco_url = REPO_URL + "data/" + coco_name + "?raw=true"

font_name = "arial.ttf"

font_url = REPO_URL + "data/" + font_name + "?raw=true"

coco_path = download_testdata(coco_url, coco_name, module="data")

font_path = download_testdata(font_url, font_name, module="data")

 

with open(coco_path) as f:

    content = f.readlines()

 

names = [x.strip() for x in content]

 

tvm.relay.testing.yolo_detection.show_detections(img, dets, thresh, names, last_layer.classes)

tvm.relay.testing.yolo_detection.draw_detections(

    font_path, img, dets, thresh, names, last_layer.classes

)

plt.imshow(img.transpose(1, 2, 0))

plt.show()

 

 

参考链接：

https://github.com/makihiro/tvm_yolov3_sample/blob/master/yolov3_quantize_sample.py

https://tvm.apache.org/docs/tutorials/frontend/from_darknet.html#sphx-glr-tutorials-frontend-from-darknet-py