训练报告: 01181215_e30_w40_fast

训练环境与参数(train)
变量名 含义
time_start 开始时间 2026-01-18 11:33:05
platform 平台 Linux-5.10.134-18.0.6.lifsea8.x86_64-x86_64-with-glibc2.35
processor CPU架构 x86_64
python_version Python版本 3.10.6
torch_version Torch版本 2.0.0+cu118
cuda_available CUDA可用 False
cuda_device_count GPU数量 0
cuda_device_name_0 GPU0名称 NA
device_used 使用设备 cpu
epochs 训练轮数 30
train_seconds 训练时长 42分钟8秒
threshold_default 默认阈值 0.5
pos_weight_used pos_weight 40.0
lr_init 初始学习率 1e-3
optimizer 优化器 Adam
report_folder_name 报告名 01181215_e30_w40
测试环境与参数(cut_test)
变量名 含义
time 记录时间 2026-01-18 12:17:06
platform 平台 Linux-5.10.134-18.0.6.lifsea8.x86_64-x86_64-with-glibc2.35
processor CPU架构 x86_64
python_version Python版本 3.10.6
torch_version Torch版本 2.0.0+cu118
cuda_available CUDA可用 False
device_used 使用设备 cpu
threshold_default 默认阈值 0.5
batch_size batch_size 64
report_folder_name 报告名 01181215_e30_w40
训练集曲线(epoch_metrics)
Loss
Precision / Recall / F1
AP(PR-AUC) / AUC(ROC)
Accuracy(若无则为空)
正类预测率 pos_pred_rate
学习率 lr(若无则为空)
建议阅读顺序:先看 pos_pred_rate 判断“乱报切点”程度,再看 PRF 是否平衡,最后结合 AP/AUC 评估整体质量。
训练集最终指标(train / val 汇总)
阈值 0.5
精确率 0.1428571428571428
召回率 0.9444444444444444
F1 0.2481751824817518
准确率 0.8860619469026548
AP(PR-AUC) 0.6972040455004045
AUC(ROC) 0.9594306496112365
TP 17
FP 102
TN 784
FN 1
正类预测率 0.1316371681415929
测试集最终指标(cut_test)
阈值 0.5
精确率 0.08695652173913043
召回率 0.7
F1 0.1546961325966851
准确率 0.8870848708487085
AP(PR-AUC) 0.7019026241023513
AUC(ROC) 0.8934831460674157
TP 14
FP 147
TN 1188
FN 6
正类预测率 0.1188191881918819
测试集帧轴可视化(每个视频:TP / FP / FN 的帧位置)
TP(预测=切点 且 GT=切点) FP(预测=切点 但 GT=非切点) FN(GT=切点 但 预测=非切点) 提示:鼠标悬停点可看帧号
V001.mp4
total_frames: 43  |  TP 1 FP 0 FN 0
GT cuts: 1 Pred cuts: 1
轴左侧≈frame 0 轴右侧≈frame 42
V002.mp4
total_frames: 161  |  TP 1 FP 2 FN 1
GT cuts: 2 Pred cuts: 3
轴左侧≈frame 0 轴右侧≈frame 160
V003.mp4
total_frames: 122  |  TP 5 FP 18 FN 0
GT cuts: 5 Pred cuts: 23
轴左侧≈frame 0 轴右侧≈frame 121
V004.mp4
total_frames: 700  |  TP 1 FP 0 FN 5
GT cuts: 6 Pred cuts: 1
轴左侧≈frame 0 轴右侧≈frame 699
V005.mp4
total_frames: 334  |  TP 6 FP 127 FN 0
GT cuts: 6 Pred cuts: 133
轴左侧≈frame 0 轴右侧≈frame 333
测试集数据明细(Excel 其它 Sheets 预览)
dataset_summary(显示前 11 行 / 共 11 行)
item value
num_videos 5
num_pairs 1355
num_cuts 20
num_non_cuts 1335
pos_ratio 0.01476014760147601
per_video_frame_stats
min_frames 43
max_frames 700
mean_frames 272
median_frames 161
per_video(显示前 5 行 / 共 5 行)
vid vid_idx total_frames gt_cut_count pred_cut_count tp fp fn gt_cuts pred_cuts
V001.mp4 0 43 1 1 1 0 0 26 26
V002.mp4 1 161 2 3 1 2 1 86,131 86,96,140
V003.mp4 2 122 5 23 5 18 0 18,35,56,72,105 18,21,25,35,36,39,40,42,43,44,45,47,48,49,51,52,53,55,56,58,61,72,105
V004.mp4 3 700 6 1 1 0 5 46,419,457,504,600,643 46
V005.mp4 4 334 6 133 6 127 0 20,69,134,169,238,263 12,13,14,15,16,17,18,19,20,31,32,35,36,44,45,46,48,54,55,57,58,59,60,61,69,70,78,79,80,97,99,100,103,110,112,113,114,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,139,140,141,142,143,148,149,151,155,156,157,159,160,166,167,168,169,173,174,176,178,179,181,183,184,186,189,191,193,196,198,199,201,203,204,206,208,209,211,213,214,219,221,223,224,226,228,229,236,238,263,265,267,272,274,275,276,277,280,281,282,284,285,286,287,289,290,291,292,294,297,299,300,311,312,314,324,325,327,329,330
classification_report(显示前 5 行 / 共 5 行)
text
precision recall f1-score support
Non-cut 0.9950 0.8899 0.9395 1335
Cut 0.0870 0.7000 0.1547 20
accuracy 0.8871 1355
下载