search

gengyanlei / fire-smoke-detect-yolov4

fire-smoke-detect-yolov4-yolov5 and fire-smoke-detection-dataset 火灾检测，烟雾检测
MIT License
1.21k stars 299 forks source link

Improving Map #47

Closed FarhanLhant closed 1 year ago

FarhanLhant commented 2 years ago

Excuse me, i already train my weight using yoloV4 configuration from alexey and some augmentation data like rotate, brightness contrast, etc and have running about 6000 iteration but only got Map around 65% when using your dataset , when i see your weight you have Map around 75% could you please tell me what data augmentation are you using or what hyperparameter and augmentation best for this case , thanks

gengyanlei commented 1 year ago

@FarhanLhant The amount of data I use is more, about 6000 images. So, the effect is better. All other parameters are default. The data I shared is the data that was marked for the first time. After collecting data continuously, and changing some marking rules, the effect will be better.

FarhanLhant commented 1 year ago

ahh I see thanks for your explanation