Closed ResonWang closed 1 year ago
YuHengsss
commented
1 year ago For the VID and OVIS datasets, it will take about 3~4 epochs to get the best result, and then the performance decreases somewhat. You can try more augmentation in training, the code could be found in our x-large model pipeline.
kangkang189
commented
1 year ago Dear author, I've tried to train your yolov on my dataset and an obvious improvement was observed almost at the first epoch of fine-turing, but in later epoches, the performance never got better. If a big lr is used, the performace will get worse quickly from the first epoch. Did you meet similar situation during the fine-turing training on pretrained model ? 你好，请问下你复现了作者YOLOV在VID数据集上的结果了吗，如果复现了能否+wechat像您咨询下如何正确的训练并得到85+MAP呢？谢谢！
ResonWang
commented
1 year ago 我只在自己的数据上实验了
WRX @.***
------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2022年12月24日(星期六) 中午11:15 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [YuHengsss/YOLOV] Is the best result from the first fine-turing epoch ? (Issue #22)
Dear author, I've tried to train your yolov on my dataset and an obvious improvement was observed almost at the first epoch of fine-turing, but in later epoches, the performance never got better. If a big lr is used, the performace will get worse quickly from the first epoch. Did you meet similar situation during the fine-turing training on pretrained model ? 你好，请问下你复现了作者YOLOV在VID数据集上的结果了吗，如果复现了能否+wechat像您咨询下如何正确的训练并得到85+MAP呢？谢谢！
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
asmallcat
commented
1 year ago 我只在自己的数据上实验了 WRX @. … ------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2022年12月24日(星期六) 中午11:15 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [YuHengsss/YOLOV] Is the best result from the first fine-turing epoch ? (Issue #22) Dear author, I've tried to train your yolov on my dataset and an obvious improvement was observed almost at the first epoch of fine-turing, but in later epoches, the performance never got better. If a big lr is used, the performace will get worse quickly from the first epoch. Did you meet similar situation during the fine-turing training on pretrained model ? 你好，请问下你复现了作者YOLOV在VID数据集上的结果了吗，如果复现了能否+wechat像您咨询下如何正确的训练并得到85+MAP呢？谢谢！ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>
您好，问一下在自己数据集上实验如何生成train_seq.npy和val_seq.npy，.npy里面存储的数据格式和内容是啥？谢谢
ResonWang
commented
1 year ago 记不太清了，作者应该提供了样本，读取看一下内容，按照原格式生成自己的
WRX @.***
------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2023年1月5日(星期四) 中午1:27 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [YuHengsss/YOLOV] Is the best result from the first fine-turing epoch ? (Issue #22)
我只在自己的数据上实验了 WRX @. … ------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2022年12月24日(星期六) 中午11:15 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [YuHengsss/YOLOV] Is the best result from the first fine-turing epoch ? (Issue #22) Dear author, I've tried to train your yolov on my dataset and an obvious improvement was observed almost at the first epoch of fine-turing, but in later epoches, the performance never got better. If a big lr is used, the performace will get worse quickly from the first epoch. Did you meet similar situation during the fine-turing training on pretrained model ? 你好，请问下你复现了作者YOLOV在VID数据集上的结果了吗，如果复现了能否+wechat像您咨询下如何正确的训练并得到85+MAP呢？谢谢！ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>
您好，问一下在自己数据集上实验如何生成train_seq.npy和val_seq.npy，.npy里面存储的数据格式和内容是啥？谢谢
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
Dear author, I've tried to train your yolov on my dataset and an obvious improvement was observed almost at the first epoch of fine-turing, but in later epoches, the performance never got better. If a big lr is used, the performace will get worse quickly from the first epoch. Did you meet similar situation during the fine-turing training on pretrained model ?