You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
222 lines
9.9 KiB
222 lines
9.9 KiB
from abc import abstractmethod,ABC
|
|
from shapely.geometry import Point, Polygon
|
|
from myutils.ConfigManager import myCongif
|
|
import numpy as np
|
|
import cv2
|
|
import ast
|
|
import platform
|
|
if myCongif.get_data("model_platform") == "acl":
|
|
import acl
|
|
|
|
#-----acl相关------
|
|
SUCCESS = 0 # 成功状态值
|
|
FAILED = 1 # 失败状态值
|
|
ACL_MEM_MALLOC_NORMAL_ONLY = 2 # 申请内存策略, 仅申请普通页
|
|
|
|
class ModelBase(ABC):
|
|
def __init__(self,path):
|
|
'''
|
|
模型类实例化
|
|
:param path: 模型文件本身的路径
|
|
:param threshold: 模型的置信阈值
|
|
'''
|
|
self.name = None #基于name来查询，用户对模型的配置参数，代表着模型名称需要唯一 2024-6-18 -逻辑还需要完善和验证
|
|
self.version = None
|
|
self.model_type = None # 模型类型 1-图像分类，2-目标检测（yolov5)，3-分割模型，4-关键点
|
|
self.system = myCongif.get_data("model_platform") #platform.system() #获取系统平台
|
|
self.do_map = { # 定义插件的入口函数 --
|
|
# POCType.POC: self.do_verify,
|
|
# POCType.SNIFFER: self.do_sniffer,
|
|
# POCType.BRUTE: self.do_brute
|
|
}
|
|
self.model_path = path # 模型路径
|
|
|
|
|
|
def __del__(self):
|
|
print("资源释放")
|
|
|
|
def draw_polygon(self, img, polygon_points,color=(0, 255, 0)):
|
|
self.polygon = Polygon(ast.literal_eval(polygon_points))
|
|
|
|
points = np.array([self.polygon.exterior.coords], dtype=np.int32)
|
|
cv2.polylines(img, points, isClosed=True, color=color, thickness=2)
|
|
|
|
def is_point_in_region(self, point):
|
|
'''判断点是否在区域内，需要先执行draw_polygon'''
|
|
if self.polygon:
|
|
return self.polygon.contains(Point(point))
|
|
else:
|
|
return False
|
|
|
|
#acl ----- 相关-----
|
|
def _init_acl(self):
|
|
'''acl初始化函数'''
|
|
self.device_id = 0
|
|
#step1 初始化
|
|
ret = acl.init()
|
|
ret = acl.rt.set_device(self.device_id) # 指定运算的Device
|
|
if ret:
|
|
raise RuntimeError(ret)
|
|
self.context, ret = acl.rt.create_context(self.device_id) # 显式创建一个Context
|
|
if ret:
|
|
raise RuntimeError(ret)
|
|
print('Init ACL Successfully')
|
|
|
|
def _del_acl(self):
|
|
'''acl去初始化'''
|
|
ret = acl.rt.destroy_context(self.context) # 释放 Context
|
|
if ret:
|
|
raise RuntimeError(ret)
|
|
ret = acl.rt.reset_device(self.device_id) # 释放Device
|
|
if ret:
|
|
raise RuntimeError(ret)
|
|
ret = acl.finalize() # 去初始化
|
|
if ret:
|
|
raise RuntimeError(ret)
|
|
print('Deinit ACL Successfully')
|
|
|
|
def _init_resource(self):
|
|
''' 初始化模型、输出相关资源。相关数据类型: aclmdlDesc aclDataBuffer aclmdlDataset'''
|
|
print("Init model resource")
|
|
# 加载模型文件
|
|
#self.model_path = "/home/HwHiAiUser/samples/yolo_acl_sample/yolov5s_bs1.om"
|
|
self.model_id, ret = acl.mdl.load_from_file(self.model_path) # 加载模型
|
|
if ret != 0:
|
|
print(f"{self.model_path}---模型加载失败！")
|
|
return False
|
|
self.model_desc = acl.mdl.create_desc() # 初始化模型信息对象
|
|
ret = acl.mdl.get_desc(self.model_desc, self.model_id) # 根据模型获取描述信息
|
|
print("[Model] Model init resource stage success")
|
|
# 创建模型输出 dataset 结构
|
|
self._gen_output_dataset() # 创建模型输出dataset结构
|
|
return True
|
|
|
|
def _gen_output_dataset(self):
|
|
''' 组织输出数据的dataset结构 '''
|
|
ret = SUCCESS
|
|
self._output_num = acl.mdl.get_num_outputs(self.model_desc) # 获取模型输出个数
|
|
self.output_dataset = acl.mdl.create_dataset() # 创建输出dataset结构
|
|
for i in range(self._output_num):
|
|
temp_buffer_size = acl.mdl.get_output_size_by_index(self.model_desc, i) # 获取模型输出个数
|
|
temp_buffer, ret = acl.rt.malloc(temp_buffer_size, ACL_MEM_MALLOC_NORMAL_ONLY) # 为每个输出申请device内存
|
|
dataset_buffer = acl.create_data_buffer(temp_buffer,
|
|
temp_buffer_size) # 创建输出的data buffer结构,将申请的内存填入data buffer
|
|
_, ret = acl.mdl.add_dataset_buffer(self.output_dataset, dataset_buffer) # 将 data buffer 加入输出dataset
|
|
|
|
if ret == FAILED:
|
|
self._release_dataset(self.output_dataset) # 失败时释放dataset
|
|
print("[Model] create model output dataset success")
|
|
|
|
def _gen_input_dataset(self, input_list):
|
|
''' 组织输入数据的dataset结构 '''
|
|
ret = SUCCESS
|
|
self._input_num = acl.mdl.get_num_inputs(self.model_desc) # 获取模型输入个数
|
|
self.input_dataset = acl.mdl.create_dataset() # 创建输入dataset结构
|
|
for i in range(self._input_num):
|
|
item = input_list[i] # 获取第 i 个输入数据
|
|
data_ptr = acl.util.bytes_to_ptr(item.tobytes()) # 获取输入数据字节流
|
|
size = item.size * item.itemsize # 获取输入数据字节数
|
|
dataset_buffer = acl.create_data_buffer(data_ptr, size) # 创建输入dataset buffer结构, 填入输入数据
|
|
_, ret = acl.mdl.add_dataset_buffer(self.input_dataset, dataset_buffer) # 将dataset buffer加入dataset
|
|
|
|
if ret == FAILED:
|
|
self._release_dataset(self.input_dataset) # 失败时释放dataset
|
|
print("[Model] create model input dataset success")
|
|
|
|
def _unpack_bytes_array(self, byte_array, shape, datatype):
|
|
''' 将内存不同类型的数据解码为numpy数组 '''
|
|
np_type = None
|
|
|
|
# 获取输出数据类型对应的numpy数组类型和解码标记
|
|
if datatype == 0: # ACL_FLOAT
|
|
np_type = np.float32
|
|
elif datatype == 1: # ACL_FLOAT16
|
|
np_type = np.float16
|
|
elif datatype == 3: # ACL_INT32
|
|
np_type = np.int32
|
|
elif datatype == 8: # ACL_UINT32
|
|
np_type = np.uint32
|
|
else:
|
|
print("unsurpport datatype ", datatype)
|
|
return
|
|
|
|
# 将解码后的数据组织为numpy数组,并设置shape和类型
|
|
return np.frombuffer(byte_array, dtype=np_type).reshape(shape)
|
|
|
|
def _output_dataset_to_numpy(self):
|
|
''' 将模型输出解码为numpy数组 '''
|
|
dataset = []
|
|
# 遍历每个输出
|
|
for i in range(self._output_num):
|
|
buffer = acl.mdl.get_dataset_buffer(self.output_dataset, i) # 从输出dataset中获取buffer
|
|
data_ptr = acl.get_data_buffer_addr(buffer) # 获取输出数据内存地址
|
|
size = acl.get_data_buffer_size(buffer) # 获取输出数据字节数
|
|
narray = acl.util.ptr_to_bytes(data_ptr, size) # 将指针转为字节流数据
|
|
|
|
# 根据模型输出的shape和数据类型,将内存数据解码为numpy数组
|
|
outret = acl.mdl.get_output_dims(self.model_desc, i)[0]
|
|
dims = outret["dims"] # 获取每个输出的维度
|
|
print(f"name:{outret['name']}")
|
|
print(f"dimCount:{outret['dimCount']}")
|
|
'''
|
|
dims = {
|
|
"name": xxx, #tensor name
|
|
"dimCount":xxx，#shape中的维度个数
|
|
"dims": [xx, xx, xx] # 维度信息 --- 取的这个
|
|
}
|
|
'''
|
|
datatype = acl.mdl.get_output_data_type(self.model_desc, i) # 获取每个输出的数据类型 --就数据类型float16,int8等
|
|
output_nparray = self._unpack_bytes_array(narray, tuple(dims), datatype) # 解码为numpy数组
|
|
dataset.append(output_nparray)
|
|
return dataset
|
|
|
|
def execute(self, input_list):
|
|
'''创建输入dataset对象, 推理完成后, 将输出数据转换为numpy格式'''
|
|
self._gen_input_dataset(input_list) # 创建模型输入dataset结构
|
|
ret = acl.mdl.execute(self.model_id, self.input_dataset, self.output_dataset) # 调用离线模型的execute推理数据
|
|
out_numpy = self._output_dataset_to_numpy() # 将推理输出的二进制数据流解码为numpy数组, 数组的shape和类型与模型输出规格一致
|
|
return out_numpy
|
|
|
|
def release(self):
|
|
''' 释放模型相关资源 '''
|
|
if self._is_released:
|
|
return
|
|
|
|
print("Model start release...")
|
|
self._release_dataset(self.input_dataset) # 释放输入数据结构
|
|
self.input_dataset = None # 将输入数据置空
|
|
self._release_dataset(self.output_dataset) # 释放输出数据结构
|
|
self.output_dataset = None # 将输出数据置空
|
|
|
|
if self.model_id:
|
|
ret = acl.mdl.unload(self.model_id) # 卸载模型
|
|
if self.model_desc:
|
|
ret = acl.mdl.destroy_desc(self.model_desc) # 释放模型描述信息
|
|
self._is_released = True
|
|
print("Model release source success")
|
|
|
|
def _release_dataset(self, dataset):
|
|
''' 释放 aclmdlDataset 类型数据 '''
|
|
if not dataset:
|
|
return
|
|
num = acl.mdl.get_dataset_num_buffers(dataset) # 获取数据集包含的buffer个数
|
|
for i in range(num):
|
|
data_buf = acl.mdl.get_dataset_buffer(dataset, i) # 获取buffer指针
|
|
if data_buf:
|
|
ret = acl.destroy_data_buffer(data_buf) # 释放buffer
|
|
ret = acl.mdl.destroy_dataset(dataset) # 销毁数据集
|
|
|
|
# @abstractmethod
|
|
# def infer(self, inputs): # 保留接口, 子类必须重写
|
|
# pass
|
|
|
|
|
|
@abstractmethod
|
|
def verify(self,image,data,isdraw=1):
|
|
'''
|
|
:param image: 需要验证的图片
|
|
:param data: select t1.model_id,t1.check_area,t1.polygon ,t2.duration_time,t2.proportion,t2.model_path
|
|
:param isdraw: 是否需要绘制线框：0-不绘制，1-绘制
|
|
:return: detections,bwarn,warntext bwarn:0-没有识别到符合要求的目标，1-没有识别到符合要求的目标。
|
|
'''
|
|
pass
|