python数据可视化——曲线图
85 浏览量
2020-12-21
02:04:51
上传
评论
收藏 119KB PDF 举报
python数据可视化数据可视化——曲线图曲线图
步骤：步骤：
1. 获取获取x坐标和坐标和y坐标（可能有多个坐标（可能有多个y坐标）的数据，注意数据的长度必须相同。坐标）的数据，注意数据的长度必须相同。
2. x坐标和每一组坐标和每一组y坐标组合，将曲线画在画布上。坐标组合，将曲线画在画布上。
先看效果：
Caption
上代码：
#coding=utf-8
import matplotlib.pyplot as plt
def txt_reader(txt_file):
iters, loss_ins, loss_cates, losses = [], [], [], [] with open(txt_file, 'r') as f:
lines = f.readlines()
for idx, line in enumerate(lines):
parts = line.rstrip("").split(",")
iters.append((idx+1)*50)
loss_in = float(parts[7].split(" ")[2])
loss_ins.append(loss_in)
loss_cate = float(parts[8].split(" ")[2])
loss_cates.append(loss_cate)
loss = float(parts[9].split(" ")[2][:-1])
losses.append(loss)
return iters, loss_ins, loss_cates, losses
if __name__ == "__main__":
# 获取数据，数据存储在列表当中，列表的长度必须相等
iters, loss_ins, loss_cates, losses = txt_reader("test.txt")
plt.title("Result Analysis")
plt.plot(iters, loss_ins, color='green', label='loss_ins')
plt.plot(iters, loss_cates, color='red', label='loss_cate')
plt.plot(iters, losses, color='blue', label='loss')
plt.legend() # 显示图例
plt.xlabel("iteration times")
plt.ylabel("loss")
plt.savefig("test.png", dpi=600)
plt.show()
test.txt数据如下：
{"mode": "train", "epoch": 1, "iter": 50, "lr": 0.00399, "time": 0.40244, "data_time": 0.0079, "memory": 3796, "loss_ins": 2.92358, "loss_cate": 0.86319, "loss": 3.78677}
{"mode": "train", "epoch": 1, "iter": 100, "lr": 0.00465, "time": 0.39973, "data_time": 0.00531, "memory": 3796, "loss_ins": 2.63875, "loss_cate": 0.30817, "loss": 2.94691}
{"mode": "train", "epoch": 1, "iter": 150, "lr": 0.00532, "time": 0.40629, "data_time": 0.00577, "memory": 3822, "loss_ins": 2.03759, "loss_cate": 0.26317, "loss": 2.30076}
资源评论
