-
Notifications
You must be signed in to change notification settings - Fork 5
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
issue2 #4
Comments
We also use voxel-size with 0.05m to pre-process the data. Please refer to the misc/cal_fcgf.py and misc/cal_fpfh.py for more details. |
Is it reasonable to use the pre-trained model you provided to run an RR of 92.81?Since I only have 24G of video memory, so numnode=5000, instead of 'all', this effect will not go from 93.53->92.81.
Message ID: ***@***.***>
|
Is it possible to increase the number of nodes from 5000 to 8000 or 10000? A smaller number of nodes could indeed degrade the final performance. |
your video memory has more 24G?
…---- Replied Message ----
| From | ***@***.***> |
| Date | 07/09/2024 11:20 |
| To | Jiang-HB/VBReg ***@***.***> |
| Cc | zyy1234q ***@***.***>,
Author ***@***.***> |
| Subject | Re: [Jiang-HB/VBReg] issue2 (Issue #4) |
Is it possible to increase the number of nodes from 5000 to 8000 or 10000? A smaller number of nodes could indeed degrade the final performance.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
I tested the model on both the V100 (32GB) and the TITAN RTX (24GB). Both tests were successful without encountering any ‘out of memory’ errors. Therefore, it’s advisable to increase the numnode value as much as possible on your machine with 24G memory. The numnode of 5000 is too small. |
all used “all”?
…---- Replied Message ----
| From | ***@***.***> |
| Date | 07/09/2024 11:31 |
| To | Jiang-HB/VBReg ***@***.***> |
| Cc | zyy1234q ***@***.***>,
Author ***@***.***> |
| Subject | Re: [Jiang-HB/VBReg] issue2 (Issue #4) |
I tested the model on both the V100 (32GB) and the TITAN RTX (24GB). Both tests were successful without encountering any ‘out of memory’ errors. Therefore, it’s advisable to increase the numnode value as much as possible on your machine with 24G memory. The numnode of 5000 is too small.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
Yes, I suggest adding an if-else statement at Line 40 in "test_3DMatch.py" to prevent this error. For example, if numnode > 10000 (determined by yourself), you can clip them to 10000 using src_keypts[:, :10000]. As such, you can constrain the scale of too large input. |
OK,thansks
…---- Replied Message ----
| From | ***@***.***> |
| Date | 07/09/2024 11:47 |
| To | Jiang-HB/VBReg ***@***.***> |
| Cc | zyy1234q ***@***.***>,
Author ***@***.***> |
| Subject | Re: [Jiang-HB/VBReg] issue2 (Issue #4) |
Yes, I suggest adding an if-else statement at Line 40 in "test_3DMatch.py" to prevent this error. For example, if numnode > 10000 (determined by yourself), you can clip them to 10000 using src_keypts[:, :10000]. As such, you can constrain the scale of too large input.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
Hello author, what do the horizontal and vertical coordinates in the diagram mean, and then, if I want to draw such a diagram how it should work
|
I noticed that you use 0.03cm voxels to downsample, and your comparison method is 0.05cm voxels downsample, is this an unfair comparison?
The text was updated successfully, but these errors were encountered: