摘要：
cuda11.2安装时和nvidia frameview sdk冲突怎么解决呀? - 知乎 (zhihu.com) 设置-应用-卸载nvidia frameview. 阅读全文
摘要：
工具网站 GitHub - lutzroeder/netron: Visualizer for neural network, deep learning, and machine learning models web在线网站 bottleneck.onnx (netron.app) 阅读全文
摘要：
import os import random import sys import cv2 import numpy as np from PIL import Image import sys_utils import nn_utils from sys_utils import _single_ 阅读全文
摘要：
因为是pixel映射到对应的M矩阵对应位置，并且只需要前4个值（left,right,top,bottom） merge_mosaic_pixel_annotations = np.arange(0, 10).reshape(2, 5) print(f"merge_mosaic_pixel_anno 阅读全文
摘要：
import json import numpy as np import torch from scipy.cluster.vq import kmeans2,kmeans from sklearn.cluster import k_means import random with open("i 阅读全文
摘要：
k=2 data = np.random.randn(100,2) data[:50]+=2.0 center_index = np.random.choice(np.arange(len(data)), k) center = data[center_index] iters = 20 def d 阅读全文
摘要：
按照面积排序 from scipy.cluster.vq import kmeans import numpy as np new_anchor = np.arange(0,9).reshape(3,3) print(new_anchor) print(new_anchor.prod(axis =1 阅读全文
摘要：
from albumentations.core.transforms_interface import DualTransform from albumentations.pytorch import ToTensorV2, ToTensor import random View Code # 随 阅读全文
摘要：
import numpy as np def draw_gauss(heatmap, x, y, gsize): height, width = heatmap.shape[:2] gsize += 1 - (gsize % 2) sigma = gsize / 6 s = 2 * sigma * 阅读全文
摘要：
class GIoULoss(nn.Module): def __init__(self): super().__init__() def forward(self, A, B): num_bbox = A.size(0) * A.size(2) ax, ay, ar, ab = A[:, 0], 阅读全文