yolov5篇---官方代码docker部署训练
作者：互联网
yolov5训练自己的数据集（docker）
yolov5官方代码： 选择 tag v5.0
https://github.com/ultralytics/yolov5
训练方式可参照另一篇博客
https://blog.csdn.net/m0_46825740/article/details/119956658?spm=1001.2014.3001.5501
数据放置路径
1. 生成docker镜像
docker build -t yolov5:5.0 .
镜像成功生成
2. 生成容器
nvidia-docker run -it -p 2224:22 -p 6006:6006 --ipc=host -v /home/slifeai/project_object/num_2/yolov5-5.0:/usrc/app --name yolov5_train yolov5:5.0 /bin/bash
容器成功生成
3. 将训练数据拷贝进容器中
docker cp litter/ 8ac6770f1edb:/usr/src/app
- 拷贝前
- 拷贝后
4. 开始训练
python train.py --data data/mydata.yaml --cfg models/yolov5s.yaml --weights 'yolov5s.pt' --batch-size 64
出现bug
File "train.py", line 543, in <module>
train(hyp, opt, device, tb_writer)
File "train.py", line 87, in train
ckpt = torch.load(weights, map_location=device) # load checkpoint
File "/opt/conda/lib/python3.8/site-packages/torch/serialization.py", line 592, in load
return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
File "/opt/conda/lib/python3.8/site-packages/torch/serialization.py", line 851, in _load
result = unpickler.load()
AttributeError: Can't get attribute 'SPPF' on <module 'models.common' from '/usr/src/app/models/common.py'>
说明model/common.py这里没有SPPF这个方法，我把yolov5-master的model/common.py里的SPPF方法拷贝进去，就能成功运行了
yolov5生成的pt权重转换
参考代码：https://github.com/soloIife/yolov5_for_rknn
数据放置路径
1. 生成docker镜像
docker build -t yolov5_for_rknn:master .
镜像成功生成
2. 生成容器
nvidia-docker run -it -p 2225:22 --ipc=host -v /home/slifeai/project_object/num_3/yolov5_for_rknn-master:/usrc/app --name yolov5_convert_weight yolov5_for_rknn:master /bin/bash
容器成功生成
3.pt转onnx
用这个容器映射到本机的这个脚本将pt转换成onnx
D:/rknn/yolov5_for_rknn-master/yolov5_original/export_no_focus.py
4.onnx转rknn
D:/rknn/rknn_convert/onnx2rknn.py ---->同D:/rknn/yolov5_for_rknn-master/yolov5_original/onnx2rknn.py
import argparse
import os
from rknn.api import RKNN
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("-i", '--onnx', type=str, default='weights/litter_10.26.onnx', help='weights path') # from yolov5/models/
parser.add_argument('--rknn', type=str, default='weights/litter_10.26.rknn', help='保存路径')
parser.add_argument("-p", '--precompile', action="store_true", help='是否是预编译模型')
parser.add_argument("-o", '--original', action="store_true", help='是否是yolov5原生的模型')
parser.add_argument("-bs", '--batch-size', type=int, default=1, help='batch size')
opt = parser.parse_args()
ONNX_MODEL = opt.onnx
if opt.rknn:
RKNN_MODEL = opt.rknn
else:
RKNN_MODEL = "%s.rknn" % os.path.splitext(ONNX_MODEL)[0]
rknn = RKNN()
print('--> config model')
rknn.config(mean_values=[[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]],
std_values=[[255.0, 255.0, 255.0, 255.0, 255.0, 255.0, 255.0, 255.0, 255.0, 255.0, 255.0, 255.0]],
batch_size=opt.batch_size, reorder_channel='0 1 2') # reorder_channel='0 1 2',
# Load tensorflow model
print('--> Loading model')
ret = rknn.load_onnx(model=ONNX_MODEL)
assert ret == 0, "Load onnx failed!"
# Build model
print('--> Building model')
if opt.precompile:
ret = rknn.build(do_quantization=True, dataset='./data/dataset1.txt', pre_compile=True) # pre_compile=True
else:
ret = rknn.build(do_quantization=True, dataset='./data/dataset1.txt')
assert ret == 0, "Build onnx failed!"
# Export rknn model
print('--> Export RKNN model')
ret = rknn.export_rknn(RKNN_MODEL)
assert ret == 0, "Export %s.rknn failed!" % opt.rknn
print('done')
5.rknn检测
D:/rknn/rknn_convert/rknn_detect.py
import cv2
import time
import random
import numpy as np
from rknn.api import RKNN
"""
yolov5 预测脚本 for rknn
"""
def get_max_scale(img, max_w, max_h):
h, w = img.shape[:2]
scale = min(max_w / w, max_h / h, 1)
return scale
def get_new_size(img, scale):
return tuple(map(int, np.array(img.shape[:2][::-1]) * scale))
def sigmoid(x):
return 1 / (1 + np.exp(-x))
def filter_boxes(boxes, box_confidences, box_class_probs, conf_thres):
box_scores = box_confidences * box_class_probs # 条件概率， 在该cell存在物体的概率的基础上是某个类别的概率
box_classes = np.argmax(box_scores, axis=-1) # 找出概率最大的类别索引
box_class_scores = np.max(box_scores, axis=-1) # 最大类别对应的概率值
pos = np.where(box_class_scores >= conf_thres) # 找出概率大于阈值的item
# pos = box_class_scores >= OBJ_THRESH # 找出概率大于阈值的item
boxes = boxes[pos]
classes = box_classes[pos]
scores = box_class_scores[pos]
return boxes, classes, scores
def nms_boxes(boxes, scores, iou_thres):
x = boxes[:, 0]
y = boxes[:, 1]
w = boxes[:, 2]
h = boxes[:, 3]
areas = w * h
order = scores.argsort()[::-1]
keep = []
while order.size > 0:
i = order[0]
keep.append(i)
xx1 = np.maximum(x[i], x[order[1:]])
yy1 = np.maximum(y[i], y[order[1:]])
xx2 = np.minimum(x[i] + w[i], x[order[1:]] + w[order[1:]])
yy2 = np.minimum(y[i] + h[i], y[order[1:]] + h[order[1:]])
w1 = np.maximum(0.0, xx2 - xx1 + 0.00001)
h1 = np.maximum(0.0, yy2 - yy1 + 0.00001)
inter = w1 * h1
ovr = inter / (areas[i] + areas[order[1:]] - inter)
inds = np.where(ovr <= iou_thres)[0]
order = order[inds + 1]
keep = np.array(keep)
return keep
def plot_one_box(x, img, color=None, label=None, line_thickness=None):
tl = line_thickness or round(0.002 * (img.shape[0] + img.shape[1]) / 2) + 1 # line/font thickness
color = color or [random.randint(0, 255) for _ in range(3)]
c1, c2 = (int(x[0]), int(x[1])), (int(x[2]), int(x[3]))
cv2.rectangle(img, c1, c2, color, thickness=tl, lineType=cv2.LINE_AA)
if label:
tf = max(tl - 1, 1) # font thickness
t_size = cv2.getTextSize(label, 0, fontScale=tl / 3, thickness=tf)[0]
c2 = c1[0] + t_size[0], c1[1] - t_size[1] - 3
cv2.rectangle(img, c1, c2, color, -1, cv2.LINE_AA) # filled
cv2.putText(img, label, (c1[0], c1[1] - 2), 0, tl / 3, [225, 255, 255], thickness=tf, lineType=cv2.LINE_AA)
def auto_resize(img, max_w, max_h):
h, w = img.shape[:2]
scale = min(max_w / w, max_h / h, 1)
new_size = tuple(map(int, np.array(img.shape[:2][::-1]) * scale))
return cv2.resize(img, new_size), scale
def letterbox(img, new_wh=(416, 416), color=(114, 114, 114)):
new_img, scale = auto_resize(img, *new_wh)
shape = new_img.shape
new_img = cv2.copyMakeBorder(new_img, 0, new_wh[1] - shape[0], 0, new_wh[0] - shape[1], cv2.BORDER_CONSTANT,
value=color)
return new_img, (new_wh[0] / scale, new_wh[1] / scale)
def load_model(model_path, npu_id):
rknn = RKNN()
devs = rknn.list_devices()
device_id_dict = {}
for index, dev_id in enumerate(devs[-1]):
if dev_id[:2] != 'TS':
device_id_dict[0] = dev_id
if dev_id[:2] == 'TS':
device_id_dict[1] = dev_id
print('-->loading model : ' + model_path)
rknn.load_rknn(model_path)
# print('--> Init runtime environment on: ' + device_id_dict[npu_id])
ret = rknn.init_runtime()
if ret != 0:
print('Init runtime environment failed')
exit(ret)
print('done')
return rknn
#
# def load_model(path, platform):
# rknn = RKNN()
# print('-->loading model')
# rknn.load_rknn(path)
# print('loading model done')
# print('--> Init runtime environment')
# # ret = rknn.init_runtime(target='rk1808', target_sub_class='AICS')
# ret = rknn.init_runtime(target=platform)
# if ret != 0:
# print('Init runtime environment failed')
# exit(ret)
# print('done')
# return rknn
class Detector:
def __init__(self, opt):
opt = opt['opt']
self.opt = opt
print(opt)
model = opt['model']
wh = opt['size']
masks = opt['masks']
anchors = opt['anchors']
names = opt['names']
conf_thres = opt['conf_thres']
iou_thres = opt['iou_thres']
platform = opt['platform']
self.wh = wh
self.size = wh
self._masks = masks
self._anchors = anchors
self.names = list(
filter(lambda a: len(a) > 0, map(lambda x: x.strip(), open(names, "r").read().split()))) if isinstance(
names, str) else names
self.conf_thres = conf_thres
self.iou_thres = iou_thres
if isinstance(model, str):
model = load_model(model, platform)
self._rknn = model
self.draw_box = False
def _predict(self, img_src, img, gain):
src_h, src_w = img_src.shape[:2]
# _img = cv2.cvtColor(_img, cv2.COLOR_BGR2RGB)
# img = img[:, :, ::-1].transpose(2, 0, 1)[None]
# # _img = np.transpose(_img[None], (0, 3, 1, 2))
# img = np.concatenate([img[..., ::2, ::2], img[..., 1::2, ::2], img[..., ::2, 1::2], img[..., 1::2, 1::2]], 1)
# img = np.transpose(img, (0, 2, 3, 1))
img = img[..., ::-1] # ？
img = np.concatenate([img[::2, ::2], img[1::2, ::2], img[::2, 1::2], img[1::2, 1::2]], 2)
t0 = time.time()
pred_onx = self._rknn.inference(inputs=[img])
print("inference time:\t", time.time() - t0)
boxes, classes, scores = [], [], []
for t in range(3):
input0_data = sigmoid(pred_onx[t][0])
input0_data = np.transpose(input0_data, (1, 2, 0, 3))
grid_h, grid_w, channel_n, predict_n = input0_data.shape
anchors = [self._anchors[i] for i in self._masks[t]]
box_confidence = input0_data[..., 4]
box_confidence = np.expand_dims(box_confidence, axis=-1)
box_class_probs = input0_data[..., 5:]
box_xy = input0_data[..., :2]
box_wh = input0_data[..., 2:4]
col = np.tile(np.arange(0, grid_w), grid_h).reshape(-1, grid_w)
row = np.tile(np.arange(0, grid_h).reshape(-1, 1), grid_w)
col = col.reshape((grid_h, grid_w, 1, 1)).repeat(3, axis=-2)
row = row.reshape((grid_h, grid_w, 1, 1)).repeat(3, axis=-2)
grid = np.concatenate((col, row), axis=-1)
box_xy = box_xy * 2 - 0.5 + grid
box_wh = (box_wh * 2) ** 2 * anchors
box_xy /= (grid_w, grid_h) # 计算原尺寸的中心
box_wh /= self.wh # 计算原尺寸的宽高
box_xy -= (box_wh / 2.) # 计算原尺寸的中心
box = np.concatenate((box_xy, box_wh), axis=-1)
res = filter_boxes(box, box_confidence, box_class_probs, self.conf_thres)
boxes.append(res[0])
classes.append(res[1])
scores.append(res[2])
boxes, classes, scores = np.concatenate(boxes), np.concatenate(classes), np.concatenate(scores)
nboxes, nclasses, nscores = [], [], []
for c in set(classes):
inds = np.where(classes == c)
b = boxes[inds]
c = classes[inds]
s = scores[inds]
keep = nms_boxes(b, s, self.iou_thres)
nboxes.append(b[keep])
nclasses.append(c[keep])
nscores.append(s[keep])
if len(nboxes) < 1:
return [], []
boxes = np.concatenate(nboxes)
classes = np.concatenate(nclasses)
scores = np.concatenate(nscores)
label_list = []
box_list = []
score_list = []
for (x, y, w, h), score, cl in zip(boxes, scores, classes):
x *= gain[0]
y *= gain[1]
w *= gain[0]
h *= gain[1]
x1 = max(0, np.floor(x).astype(int))
y1 = max(0, np.floor(y).astype(int))
x2 = min(src_w, np.floor(x + w + 0.5).astype(int))
y2 = min(src_h, np.floor(y + h + 0.5).astype(int))
# label_list.append(self.names[cl])
label_list.append(cl)
score = round(score, 3)
score_list.append(score)
box_list.append((x1, y1, x2, y2))
if self.draw_box:
plot_one_box((x1, y1, x2, y2), img_src, label=self.names[cl])
print("label_list", label_list)
print("score_list", score_list)
print("box_list", box_list)
return label_list, np.array(box_list)
def detect_resize(self, img_src):
"""
预测一张图片，预处理使用resize
return: labels,boxes
"""
_img = cv2.resize(img_src, self.wh)
gain = img_src.shape[:2][::-1]
return self._predict(img_src, _img, gain)
def detect(self, img_src):
"""
预测一张图片，预处理保持宽高比
return: labels,boxes
"""
_img, gain = letterbox(img_src, self.wh)
return self._predict(img_src, _img, gain)
def close(self):
self._rknn.release()
def __enter__(self):
return self
def __exit__(self, exc_type, exc_val, exc_tb):
self.close()
def __del__(self):
self.close()
# def test_video(det, video_path):
# reader = cv2.VideoCapture()
# reader.open(video_path)
# while True:
# ret, frame = reader.read()
# if not ret:
# break
# t0 = time.time()
# det.detect(frame)
# print("total time", time.time() - t0)
# cv2.imshow("res", auto_resize(frame, 1200, 600)[0])
# cv2.waitKey(1)
if __name__ == '__main__':
import yaml
import cv2
image = cv2.imread("img/0625_Bin_046.jpg")
with open("yolov5_rknn_640x640.yaml", "rb") as f:
cfg = yaml.load(f, yaml.FullLoader)
d = Detector(cfg)
d.draw_box = True
d.detect(image)
# cv2.imshow("res", image)
# cv2.waitKey()
# cv2.destroyAllWindows()
标签：box,rknn,yolov5,img,opt,self,---,np,docker 来源： https://blog.csdn.net/m0_46825740/article/details/120946942