sun15684712291/aicheckv2
1
1
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

查询训练报告接口

Browse Source
This commit is contained in:
sunyg
2025-03-11 11:42:09 +08:00
parent 7d736c4ac4
commit f49a6caf10
10 changed files with 143 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
app/service/project_detect_service.py
View File

@ -11,6 +11,7 @@ from app.config.config_reader import detect_url
from app.util import os_utils as os
from app.util import random_utils as ru
from app.config.config_reader import yolo_url
from app.websocket.web_socket_server import room_manager
def add_detect(detect_in: ProjectDetectIn, session: Session):
@ -133,14 +134,12 @@ def run_commend(weights: str, source: str, project: str, name: str,
:return:
"""
yolo_path = os.file_path(yolo_url, 'detect.py')
yield f"stdout: 模型推理开始，请稍等。。。 \n"
room = 'detect_' + str(detect_id)
await room_manager.send_to_room(room, f"stdout: 模型训练开始，请稍等。。。\n")
commend = ["python", '-u', yolo_path, "--weights", weights, "--source", source, "--name", name, "--project", project, "--save-txt"]
# 启动子进程
with subprocess.Popen(
["python", '-u', yolo_path,
"--weights", weights,
"--source", source,
"--name", name,
"--project", project],
commend,
bufsize=1, # bufsize=0时，为不缓存；bufsize=1时，按行缓存；bufsize为其他正整数时，为按照近似该正整数的字节数缓存
shell=False,
stdout=subprocess.PIPE,
@ -152,13 +151,14 @@ def run_commend(weights: str, source: str, project: str, name: str,
line = process.stdout.readline()
process.stdout.flush() # 刷新缓存，防止缓存过多造成卡死
if line != '\n':
yield line
await room_manager.send_to_room(room, line + '\n')
# 等待进程结束并获取返回码
return_code = process.wait()
if return_code != 0:
pdc.update_detect_status(detect_id, -1, session)
else:
await room_manager.send_to_room(room, 'success')
pdc.update_detect_status(detect_id, 2, session)
detect_imgs = pdc.get_img_list(detect_id, session)
detect_log_imgs = []

93
app/service/project_train_service.py
View File

@ -11,13 +11,13 @@ from app.util import os_utils as os
from app.util import random_utils as ru
from app.config.config_reader import datasets_url, runs_url, images_url, yolo_url
from app.websocket.web_socket_server import room_manager
from app.util.csv_utils import read_csv
from sqlalchemy.orm import Session
from typing import List
from fastapi import UploadFile
import yaml
import subprocess
import asyncio
def add_project(info: ProjectInfoIn, session: Session, user_id: int):
@ -221,7 +221,7 @@ async def run_commend(data: str, project: str,
line = process.stdout.readline()
process.stdout.flush() # 刷新缓存，防止缓存过多造成卡死
if line != '\n':
await room_manager.send_to_room(room, line)
await room_manager.send_to_room(room, line + '\n')
# 等待进程结束并获取返回码
return_code = process.wait()
@ -234,8 +234,10 @@ async def run_commend(data: str, project: str,
train = ProjectTrain()
train.project_id = project_id
train.train_version = name
bast_pt_path = os.file_path(project, name, 'weights', 'best.pt')
last_pt_path = os.file_path(project, name, 'weights', 'last.pt')
train_url = os.file_path(project, name)
train.train_url = train_url
bast_pt_path = os.file_path(train_url, 'weights', 'best.pt')
last_pt_path = os.file_path(train_url, 'weights', 'last.pt')
train.best_pt = bast_pt_path
train.last_pt = last_pt_path
if weights != None and weights != '':
@ -275,19 +277,76 @@ def operate_img_label(img_list: List[ProjectImgLabel],
+ image_label.mark_height + '\n')
def split_array(data):
def get_train_result(train_id: int, session: Session):
"""
将数组按照3:1的比例切分成两个数组。
:param data: 原始数组
:return: 按照3:1比例分割后的两个数组
根据result.csv文件查询训练报告
:param train_id:
:param session:
:return:
"""
total_length = len(data)
if total_length < 4:
raise ValueError("数组长度至少需要为4才能进行3:1的分割")
# 计算分割点
split_index = (total_length * 3) // 4
# 使用切片分割数组
part1 = data[:split_index]
part2 = data[split_index:]
return part1, part2
train_info = ptc.get_train(train_id, session)
if train_info is None:
return None
result_csv_path = os.file_path(train_info.train_url, 'results.csv')
result_row = read_csv(result_csv_path)
report_data = {}
# 轮数
epoch_data = []
# 边界框回归损失（Bounding Box Loss），衡量预测框位置（中心坐标、宽高）与真实框的差异，值越低表示定位越准。
train_box_loss = []
# 目标置信度损失（Objectness Loss），衡量检测到目标的置信度误差（即是否包含物体），值越低表示模型越能正确判断有无物体。
train_obj_loss = []
# 分类损失（Classification Loss），衡量预测类别与真实类别的差异，值越低表示分类越准。
train_cls_loss = []
# 验证集的边界框回归损失，反映模型在未见数据上的定位能力。
val_box_loss = []
# 验证集的目标置信度损失，反映模型在未见数据上判断物体存在的能力。
val_obj_loss = []
# 验证集的分类损失，反映模型在未见数据上的分类准确性。
val_cls_loss = []
# 精确率（Precision）：正确检测的正样本占所有预测为正样本的比例，反映“误检率”。值越高说明误检越少。
m_p = []
# 召回率（Recall）：正确检测的正样本占所有真实正样本的比例，反映“漏检率”。值越高说明漏检越少。
m_r = []
# 主干网络（Backbone）的学习率。
x_lr0 = []
# 检测头（Head）的学习率。
x_lr1 = []
for row in result_row:
epoch_data.append(row[0].strip())
train_box_loss.append(row[1].strip())
train_obj_loss.append(row[2].strip())
train_cls_loss.append(row[3].strip())
val_box_loss.append(row[8].strip())
val_obj_loss.append(row[9].strip())
val_cls_loss.append(row[10].strip())
m_p.append(row[4].strip())
m_r.append(row[5].strip())
x_lr0.append(row[11].strip())
x_lr1.append(row[12].strip())
report_data['epoch_data'] = epoch_data
report_data['train_box_loss'] = train_box_loss
report_data['train_obj_loss'] = train_obj_loss
report_data['train_cls_loss'] = train_cls_loss
report_data['val_box_loss'] = val_box_loss
report_data['val_obj_loss'] = val_obj_loss
report_data['val_cls_loss'] = val_cls_loss
report_data['m_p'] = m_p
report_data['m_r'] = m_r
report_data['x_lr0'] = x_lr0
report_data['x_lr1'] = x_lr1
return report_data