-
Notifications
You must be signed in to change notification settings - Fork 9
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Strange results #1
Comments
Hello,
You might need to do some longer training. I included a screenshot of my tensorboard logs from the traffic sign recognition tasks.
[cid:528897b4-b536-44ab-a1ef-60f9cfe7de65]
I hope you get it too work! If you have any questions in the future, feel free to contact me.
Kind regards,
Joris
…________________________________
Van: yousef-xailient <[email protected]>
Verzonden: zaterdag 1 augustus 2020 07:29
Aan: sara-nl/attention-sampling-pytorch <[email protected]>
CC: Subscribed <[email protected]>
Onderwerp: [sara-nl/attention-sampling-pytorch] Strange results (#1)
Hi there,
Thank you for providing a PyTorch implementation of the original!
I have cloned your repo and tried to run the experiment on the speed limit sign dataset and I am getting strange results not close to what is shown in the paper.
I have merely cloned the repo, ran your speed limit dataset script to download the data, and speed_limit.py to run the job.
Any help would be appreciated!
[Screenshot from 2020-08-01 15-26-54]<https://user-images.githubusercontent.com/57377547/89094841-a408f280-d40b-11ea-89df-18b6465bbbaf.png>
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub<#1>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADYARHALR6WKE5NO4TQLKX3R6OR2JANCNFSM4PRODCWQ>.
|
Hi, I am experimenting the same issue here. I tried MegaMNIST with no problem but had the same problem on speed limit dataset. I can not see the screenshot of your tensorboard logs. Could you please share it or tell me how many epochs is required? I really appreciate it. |
Hi,
For the speed limit dataset I am running for a 1000 epochs. You can download the tensorboard log file [here][https://we.tl/t-Sf0IrRDcJJ].
Kind regards,
Joris
…________________________________
Van: Jingwei Zhang <[email protected]>
Verzonden: woensdag 30 september 2020 06:36
Aan: sara-nl/attention-sampling-pytorch <[email protected]>
CC: jorism1993 <[email protected]>; Comment <[email protected]>
Onderwerp: Re: [sara-nl/attention-sampling-pytorch] Strange results (#1)
Hi,
I am experimenting the same issue here. I tried MegaMNIST with no problem but had the same problem on speed limit dataset. I can not see the screenshot of your tensorboard logs. Could you please share it or tell me how many epochs is required? I really appreciate it.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub<#1 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADYARHFLO3I7M5SWERI4G43SIKYVPANCNFSM4PRODCWQ>.
|
Thank you! I'll have a try again. |
@skykiny I have a similar problem! Did you get better results with increasing the number of epochs? |
Can I ask what is the best accuracy that you get? I cannot access the log file. |
Hi there,
Thank you for providing a PyTorch implementation of the original!
I have cloned your repo and tried to run the experiment on the speed limit sign dataset and I am getting strange results not close to what is shown in the paper.
I have merely cloned the repo, ran your speed limit dataset script to download the data, and speed_limit.py to run the job.
Any help would be appreciated!
The text was updated successfully, but these errors were encountered: