help #37
Comments
|
Hi, what exactly is the error reported. You should check if the data is correctly entered into the '.npy' file, check issue 6 for that. |
|
刘明
***@***.***
#----------Creating logger----------#
#----------GPU init----------#
#----------Preparing dataset----------#
#----------Prepareing Models----------#
SC_Att_Bridge was used
#----------Prepareing loss, opt, sch and amp----------#
#----------Set other params----------#
#----------Training----------#
Traceback (most recent call last):
File "train.py", line 189, in <module>
main(config)
File "train.py", line 132, in main
train_one_epoch(
File "/share/home/ncu_418000230030/pythonProject/UltraLight-VM-UNet/engine.py", line 40, in train_one_epoch
loss.backward()
File "/public/home/ncu_418000230030/anaconda3/envs/vmunet/lib/python3.8/site-packages/torch/_tensor.py", line 487, in backward
torch.autograd.backward(
File "/public/home/ncu_418000230030/anaconda3/envs/vmunet/lib/python3.8/site-packages/torch/autograd/__init__.py", line 197, in backward
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
RuntimeError: Unable to find a valid cuDNN algorithm to run convolution
报这个错，不知道怎么解决，然后请问我上个邮件和你说的步骤是正确的吗
…------------------ 原始邮件 ------------------
发件人: "wurenkai/UltraLight-VM-UNet" ***@***.***>;
发送时间: 2024年5月16日(星期四) 凌晨1:29
***@***.***>;
***@***.******@***.***>;
主题: Re: [wurenkai/UltraLight-VM-UNet] help (Issue #37)
Hi, what exactly is the error reported. You should check if the data is correctly entered into the '.npy' file, check issue 6 for that.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
|
There is no problem with the steps. However, the problem you are currently reporting is still a data preprocessing issue. You need to check if the file directory and the preprocessing code 'Prepare_.ISIC2017py' correspond. This will need to be adjusted by you. The code is set up with a print function, and usually the data is correctly entered into the '.npy' file, which will display the processing sequence numbers one by one. |
|
已经成功跑通了，谢谢最近的一些回复
刘明
***@***.***
…------------------ 原始邮件 ------------------
发件人: "wurenkai/UltraLight-VM-UNet" ***@***.***>;
发送时间: 2024年5月16日(星期四) 下午5:53
***@***.***>;
***@***.******@***.***>;
主题: Re: [wurenkai/UltraLight-VM-UNet] help (Issue #37)
There is no problem with the steps. However, the problem you are currently reporting is still a data preprocessing issue. You need to check if the file directory and the preprocessing code 'Prepare_.ISIC2017py' correspond. This will need to be adjusted by you. The code is set up with a print function, and usually the data is correctly entered into the '.npy' file, which will display the processing sequence numbers one by one.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
请问您能复现到作者论文里的miou dice等效果吗？我一直复现不到那么高的效果 |
|
|
您有没有在ISIC2017数据集上复现的结果？结果怎样？
…---- Replied Message ----
| From | ***@***.***> |
| Date | 05/20/2024 09:45 |
| To | wurenkai/UltraLight-VM-UNet ***@***.***> |
| Cc | Saul62 ***@***.***>,
Comment ***@***.***> |
| Subject | Re: [wurenkai/UltraLight-VM-UNet] help (Issue #37) |
已经成功跑通了，谢谢最近的一些回复 刘明 @.***
…
------------------ 原始邮件 ------------------ 发件人: "wurenkai/UltraLight-VM-UNet" @.>; 发送时间: 2024年5月16日(星期四) 下午5:53 _@**._>; ***@***.***@.>; 主题: Re: [wurenkai/UltraLight-VM-UNet] help (Issue #37) There is no problem with the steps. However, the problem you are currently reporting is still a data preprocessing issue. You need to check if the file directory and the preprocessing code 'Prepare.ISIC2017py' correspond. This will need to be adjusted by you. The code is set up with a print function, and usually the data is correctly entered into the '.npy' file, which will display the processing sequence numbers one by one. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: _@_.*>
请问您能复现到作者论文里的miou dice等效果吗？我一直复现不到那么高的效果
image.png (view on web)
这是我在ISIC2018数据集上复现的结果，差距好像在可接受范围之内吧
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.Message ID: ***@***.***>
|
|
我没有在ISIC2017数据集上复现，因为我没有下载这个数据集
…------------------ 原始邮件 ------------------
发件人: "wurenkai/UltraLight-VM-UNet" ***@***.***>;
发送时间: 2024年5月20日(星期一) 上午10:11
***@***.***>;
***@***.******@***.***>;
主题: Re: [wurenkai/UltraLight-VM-UNet] help (Issue #37)
您有没有在ISIC2017数据集上复现的结果？结果怎样？
---- Replied Message ----
| From | ***@***.***> |
| Date | 05/20/2024 09:45 |
| To | wurenkai/UltraLight-VM-UNet ***@***.***> |
| Cc | Saul62 ***@***.***>,
Comment ***@***.***> |
| Subject | Re: [wurenkai/UltraLight-VM-UNet] help (Issue #37) |
已经成功跑通了，谢谢最近的一些回复 刘明 @.***
…
------------------ 原始邮件 ------------------ 发件人: "wurenkai/UltraLight-VM-UNet" @.>; 发送时间: 2024年5月16日(星期四) 下午5:53 _@**._>; ***@***@***.***>; 主题: Re: [wurenkai/UltraLight-VM-UNet] help (Issue #37) There is no problem with the steps. However, the problem you are currently reporting is still a data preprocessing issue. You need to check if the file directory and the preprocessing code 'Prepare.ISIC2017py' correspond. This will need to be adjusted by you. The code is set up with a print function, and usually the data is correctly entered into the '.npy' file, which will display the processing sequence numbers one by one. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: _@_.*>
请问您能复现到作者论文里的miou dice等效果吗？我一直复现不到那么高的效果
image.png (view on web)
这是我在ISIC2018数据集上复现的结果，差距好像在可接受范围之内吧
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.Message ID: ***@***.***>
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.Message ID: ***@***.***>
|
请问你在ISIC2018数据集上跑出来的结果有随机性吗？即使在随机种子固定的情况下 |
你好我和你出现了同样的报错，请问你是怎么解决的呢 |
是下载数据集成这样子
然后python Prepare_ISIC2017.py得到这些.npy文件
之后修改配置文件中的这里
然后就python train.py，请问我这样的顺序错了吗，跑起代码来一直报错
The text was updated successfully, but these errors were encountered: