训练问题 - Githubissues

1632325673 commented 2 years ago

你好，训练出现这种情况是哪里出了问题吗 epochs is 1000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 2000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 3000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 4000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 5000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 6000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 7000, mean_psnr is 0.000000, mean_ssim is 0.000000

hellloxiaotian commented 2 years ago

你好:

需要自己下载测试集，然后把测试集放在dataset文件夹下就可以。

祝好！田春伟副教授西北工业大学

在 2022-06-29 10:28:24，"1632325673" @.***> 写道：

你好，训练出现这种情况是哪里出了问题吗 epochs is 1000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 2000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 3000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 4000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 5000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 6000, mean_psnr is 0.000000, mean_ssim is 0.000000 epochs is 7000, mean_psnr is 0.000000, mean_ssim is 0.000000

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>

1632325673 commented 2 years ago

您好，边训练边测试的话，结果取最后一个epoch还是可以取结果最好的一次epoch

Ralapfranco commented 2 years ago

hello! can you tell me how to set epochs in the code?

hellloxiaotian commented 2 years ago

Our paper has given steps rather than epochs in this paper.

Best wishes, Chunwei

At 2022-07-24 21:17:14, "Ralap" @.***> wrote:

hello! can you tell me how to set epochs in the code?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

Ralapfranco commented 2 years ago

Our paper has given steps rather than epochs in this paper. Best wishes, Chunwei At 2022-07-24 21:17:14, "Ralap" @.> wrote: hello! can you tell me how to set epochs in the code? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.>

ok! thank you!

Ralapfranco commented 2 years ago

Our paper has given steps rather than epochs in this paper. Best wishes, Chunwei At 2022-07-24 21:17:14, "Ralap" @.> wrote: hello! can you tell me how to set epochs in the code? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.>

hello！how to choose the model after training to use test the data？ there are lots of train model.

Ralapfranco commented 2 years ago

您好，边训练边测试的话，结果取最后一个epoch还是可以取结果最好的一次epoch

so have you get it？ i have same question.

hellloxiaotian commented 2 years ago

为了保证实验的公平性，我们取得最后一个epoch.

在 2022-08-11 10:21:51，"Ralap" @.***> 写道：

您好，边训练边测试的话，结果取最后一个epoch还是可以取结果最好的一次epoch

so have you get it？ i have same question.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

Ralapfranco commented 2 years ago

尊敬的田老师，您好！谢谢您！

---- 回复的原邮件 ---- | 发件人 | Chunwei @.> | | 日期 | 2022年08月11日 12:05 | | 收件人 | @.> | | 抄送至 | @.**@.> | | 主题 | Re: [hellloxiaotian/ESRGCNN] 训练问题 (Issue #6) |

为了保证实验的公平性，我们取得最后一个epoch.

在 2022-08-11 10:21:51，"Ralap" @.***> 写道：

您好，边训练边测试的话，结果取最后一个epoch还是可以取结果最好的一次epoch

so have you get it？ i have same question.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

hellloxiaotian / ESRGCNN

训练问题 #6