Segfault during model training and bad results

[Morpheus3000]

I tried running the training on about 110k images. I used all the default parameters for the training. But after around 11 epochs, I get a segfault in the training. I don't know where this error is originating from.

Also, since I had 11 epochs already saved, I just tried running a test on the sample images provided to see the performance. But the results seems very bad. Did you notice something similar? Is it because of the epoch that I used and it can be resolved with using a higher epoch?

Along the same lines, what would you say as the optimum epoch for the training? The default is about 30 epochs.

Thank you for your time.

[zhnyuren]

I tried running the training on about 110k images. I used all the default parameters for the training. But after around 11 epochs, I get a segfault in the training. I don't know where this error is originating from.

Also, since I had 11 epochs already saved, I just tried running a test on the sample images provided to see the performance. But the results seems very bad. Did you notice something similar? Is it because of the epoch that I used and it can be resolved with using a higher epoch?

Along the same lines, what would you say as the optimum epoch for the training? The default is about 30 epochs.

Thank you for your time.

Hello, I am wondering what dataset did you use.
Thanks.

[Morpheus3000]

I tried running the training on about 110k images. I used all the default parameters for the training. But after around 11 epochs, I get a segfault in the training. I don't know where this error is originating from.
Also, since I had 11 epochs already saved, I just tried running a test on the sample images provided to see the performance. But the results seems very bad. Did you notice something similar? Is it because of the epoch that I used and it can be resolved with using a higher epoch?
Along the same lines, what would you say as the optimum epoch for the training? The default is about 30 epochs.

Thank you for your time.

Hello, I am wondering what dataset did you use.
Thanks.

I used the same dataset as mentioned in the paper, the EmotioNet dataset.

[zhnyuren]

I tried running the training on about 110k images. I used all the default parameters for the training. But after around 11 epochs, I get a segfault in the training. I don't know where this error is originating from.
Also, since I had 11 epochs already saved, I just tried running a test on the sample images provided to see the performance. But the results seems very bad. Did you notice something similar? Is it because of the epoch that I used and it can be resolved with using a higher epoch?
Along the same lines, what would you say as the optimum epoch for the training? The default is about 30 epochs.

Thank you for your time.

Hello, I am wondering what dataset did you use.
Thanks.

I used the same dataset as mentioned in the paper, the EmotioNet dataset.

thanks.

[albertpumarola]

Hi, it is hard to know what is happening to you... Make sure that all the launch process is correct and the dataset is properly preprocessed.

[ZHANG-SHI-CHANG]

I tried running the training on about 110k images. I used all the default parameters for the training. But after around 11 epochs, I get a segfault in the training. I don't know where this error is originating from.

Also, since I had 11 epochs already saved, I just tried running a test on the sample images provided to see the performance. But the results seems very bad. Did you notice something similar? Is it because of the epoch that I used and it can be resolved with using a higher epoch?

Along the same lines, what would you say as the optimum epoch for the training? The default is about 30 epochs.

Thank you for your time.

Hi, can you share the dataset with me? My application for the data set failed...my email is 18354226160@163.com, thank you very much^_^

[EricGzq]

I tried running the training on about 110k images. I used all the default parameters for the training. But after around 11 epochs, I get a segfault in the training. I don't know where this error is originating from.
Also, since I had 11 epochs already saved, I just tried running a test on the sample images provided to see the performance. But the results seems very bad. Did you notice something similar? Is it because of the epoch that I used and it can be resolved with using a higher epoch?
Along the same lines, what would you say as the optimum epoch for the training? The default is about 30 epochs.

Thank you for your time.

Hi, can you share the dataset with me? My application for the data set failed...my email is 18354226160@163.com, thank you very much^_^

Hi! Did you get this dataset? I also need one, thank you very much. My email is gzq_mail@163.com

[albertpumarola]

Closed for inactivity of the question. New posts are only asking for the dataset without the EmotioNet author's permission.